date: Sun, 26 Nov 2006 06:17:40 -0800 x-mimeole: Produced By Microsoft MimeOLE V6.00.2900.2962 from: subject: Information Theory content-type: multipart/related; type="text/html"; boundary="----=_NextPart_000_0000_01C71122.93BDACB0" mime-version: 1.0 This is a multi-part message in MIME format. ------=_NextPart_000_0000_01C71122.93BDACB0 content-type: text/html; charset="iso-8859-1" content-transfer-encoding: quoted-printable content-location: file://C:\Documents and Settings\Klaus Holtz\Desktop\infotheo.htm Information Theory Webpage

Autosophy Info= rmation=20 Theory provides lossless data and video compression based on the data=20 content

=20

Omni Dimensional=20 Networks,
Tel. Fax 415 474 - 4860,
Email:=20 holtzk@autosophy.com
=20

ABSTRACT<= /FONT>=20

A new Autosophy in= formation=20 theory provides an alternative to the classical Shannon information theory= .=20 Using the new theory in communication networks provides both a high degree= of=20 lossless compression and virtually unbreakable encryption codes for networ= k=20 security. The bandwidth in a conventional Shannon communication is determi= ned=20 only by the data volume and the hardware parameters, such as image size;=20 resolution; or frame rates in television. The data content, or what is sho= wn on=20 the screen, is irrelevant. In contrast, the bandwidth in Autosophy communi= cation=20 is determined only by data content, such as novelty and movement in televi= sion=20 images. It is the data volume and hardware parameters that become=20 irrelevant.=20

Basically, the new= =20 communication methods use prior “knowledge” of the data, store= d in a library, to=20 encode subsequent transmissions. The more “knowledge” stored i= n the libraries,=20 the higher the potential compression ratio. “Information” is r= edefined as that=20 which is not already known by the receiver. Everything already known is=20 redundant and need not be re-transmitted. In a perfect communication each= =20 transmission code, called a “Tip”, creates a new “engram= ” of knowledge in the=20 library in which each tip transmission can represent any amount of data.=20 Autosophy theories provide six separate learning modes, or Omni Dimensiona= l=20 Networks, all of which can be used for data compression. The new informati= on=20 theory reveals the theoretical flaws of other data compression methods,=20 including: the Huffman; Ziv Lempel; LZW codes and commercial compression c= odes=20 such as V.42bis and MPEG-2.=20

INTRODUCTION=20

Philosophers, rath= er than=20 engineers, have long pondered elusive concepts like “knowledge”= ;, “meaning” and=20 “information”. Because these concepts are both subjective and = intangible they=20 were thought to be outside of normal scientific research. Until quite rece= ntly=20 “knowledge” could not be defined in precise mathematical terms= for use in=20 electronic devices. There was to be no place for “meaning” in = science, computing=20 or communications.=20

In 1948 Claude Sha= nnon=20 conformed to the prevailing traditions in “A mathematical theory of= =20 communications” by declaring that “information” has no &= #8220;meaning” or that=20 “meaning” is irrelevant to the engineering problem. He argued = that any type of=20 data can be communicated by first converting it into binary digits called= =20 “bits”. The binary bits represent “information” eq= uivalent to simple Yes-No=20 answers to hypothetical questions. The only purpose of the communication i= s to=20 restore the original data pattern in the receiver. Since information ̶= 0;removes=20 uncertainty”, the more bits being transmitted, the more accurate the= =20 reconstructed data pattern in the receiver. Any attempt to remove bits in = the=20 transmission, through data compression, must “increase uncertainty&#= 8221; and=20 therefore result in inevitable data distortion or loss of resolution. The= =20 transmission bandwidth is determined only by the data volume and hardware= =20 parameters. The meaning of the data is irrelevant. According to Shannon th= e=20 “information” contained in this paper, for example, can be mea= sured by counting=20 the number of text characters and multiplying them by an 8 bit ASCII code.= =20 Whether this paper contains random number sequences, nothing but blank spa= ces,=20 or written text is irrelevant.=20

Shannon’s me= chanistic=20 information theory now totally dominates modern data communications and=20 computing.=20

Programmed data pr= ocessing=20 computers manipulate “meaningless” data patterns using arithme= tic and=20 programming, while communication networks only transmit “meaningless= ” binary bit=20 streams. And yet the following questions remain:=20

1) If Shannon̵= 7;s binary bit=20 streams truly represent “communications”, then why is communic= ation in nature or=20 between humans so entirely different? Normal communications use symbols, s= ounds,=20 or images which have “meaning”. Other than artificial computer= communications,=20 almost no natural communication is done with binary bits or yes-no=20 answers.=20

2) Isn’t = 220;Meaningless=20 information” a contradiction in terms ?. If a message contains no &#= 8220;meaning” then=20 it is of no value to the receiver and represents only gibberish. Informati= on not=20 only contains “meaning” but is “meaning”. Binary b= it streams have no “meaning”=20 to human beings and therefore contain no true information.=20

3) Why is Shannon&= #8217;s theory=20 so totally unable to explain the functioning and learning processes of our= own=20 brain? Our brain obviously does not function like a programmed data proces= sing=20 computer. It extracts “meaning” from environmental data and ac= cumulates=20 “knowledge”, without requiring any external programming or=20 arithmetic.=20

4) Why is “l= ossless” data=20 compression not only possible but already the basis of a huge market? The= =20 V.42bis compression standard, for example, is now implemented in virtually= all=20 new modems. Data compression utilities are being used in most computers to= =20 double or even triple storage capacities in disc drives. Data compression = is=20 rapidly invading all forms of communications so that uncompressed data=20 transmissions may soon become uncompetitive.=20

Enter “Autos= ophy”, an=20 emerging science initially founded to explain “self-assembling struc= tures”, such=20 as: crystals, living trees, or societies in mathematical terms. Autosophy = is a=20 combination of two Greek words “autos” (self) and “sophi= a” (knowledge or wisdom)=20 which together can be translated as “self-learning” or the und= erstanding of=20 oneself. Since 1974 this research has produced a mathematical theory of=20 “learning” and a new “information theory” which of= fers an alternative to the=20 classical Shannon information theory. Autosophy re-admits “meaning&#= 8221; and=20 “information entities” back into science, but only as mathemat= ical abstracts and=20 engineering tools. It may open the way towards an entirely different=20 communications technology and eventually to the design of self-learning=20 brain-like “Autosopher”.=20

Comparing conventi= onal data=20 processing with Autosophy learning is like comparing manufacturing and far= ming.=20 In a manufacturing plant, products are built from input materials using hu= man or=20 machine labor according to engineering design and planning. Every step in = the=20 process is under human control and supervision. Data processing computers= =20 similarly produce useful output data by combining input data using arithme= tic=20 and programming. Every step in the process is defined by a human programme= r.=20 Farming, in contrast, is done by depositing seeds into the soil and waitin= g for=20 the results. The seeds will attract selected materials from a random envir= onment=20 to assemble themselves into very complex structures, such as crystals or l= iving=20 trees. The process is fully autonomous and does not require any human desi= gn or=20 supervision. Similarly, an “autosopher” will absorb selected d= ata from a random=20 environment to grow self-learning data networks in electronic memories, i.= e.,=20 “data crystals” or “data trees”. There is no need = for human supervision or=20 programming.=20

Self-assembling st= ructures=20 violate important laws of physics, especially the laws of entropy. Accordi= ng to=20 classical physics, complex structures cannot assemble themselves without o= utside=20 guidance or control. This is because conventional science only allows=20 statistical randomness to explain the behavior of objects. Objects or part= icles=20 are not supposed to have choices, directions or a will. And yet living tre= es do=20 assemble themselves from random environmental molecules and more advanced= =20 creatures do evolve from more primitive creatures against the laws of entr= opy.=20 Physics equations usually contain only the three elements: matter, forces = and=20 dimensions. Autosophy tries to solve the enigma by introducing “info= rmation” as=20 a fourth element in the cosmos. Without “information” (genetic= codes), living=20 creatures cannot exist and entropy could only increase in the universe. Ad= ding=20 “information” (not the Shannon type information) to physics, c= hemistry and=20 biology may result in a revolution in most of these natural sciences. It t= urns=20 out that there are indeed “information entities” in most physi= cal structures and=20 there really is “meaning” in communications.=20

SHANNON VS. AUT= OSOPHY=20 COMMUNICATIONS=20

Shannon and Autoso= phy=20 communications are entirely different, following entirely different natura= l=20 laws. But, both achieve the same final result, the transmission of data or= video=20 in communication networks.=20

=20

Figure 1. A typica= l Shannon=20 data and video communication=20

The choice of whic= h method=20 to use depends on the data type and the application. Shannon communication= s are=20 normally superior for meaningless or totally random data sequences, while= =20 Autosophy communications are highly superior for meaningful data, such as= =20 written text, speech and television. Because most communications networks = are=20 built to serve human communications needs, most network communications can= be=20 greatly improved by using Autosophy methods.=20

As shown in Fig. 1= , a=20 Shannon television transmission depends only on the data volume and hardwa= re=20 parameters. The images are formed by pixels in ordered rows and columns on= the=20 screen where each pixel must be constantly re-scanned and described in the= =20 transmission. The bandwidth (bits per second) in a digital television syst= em can=20 be calculated as the product of rows, columns, brightness resolution (bits= per=20 color), frame rate (frames per second) and colors (red green blue) in colo= r=20 television. All of these values are hardware “quantities” depe= nding on the=20 physical parameter in the system. The image actually shown on the screen i= s=20 irrelevant, such that a totally random noise pattern would require just as= many=20 bits per second as a blank screen. Any increase in image size, resolution = or=20 scanning rates would require a much larger transmission bandwidth. Any att= empt=20 to remove bits, through data compression, leads to inevitable image distor= tions=20 or reduced resolution.<= /FONT>=20

In the theoretical= system=20 the input data is always regarded as “quantities” even for ASC= II text=20 transmissions. The input data is converted using Shannon’s equations= into binary=20 digits or bits. The only purpose of the communication is to reproduce the = input=20 data pattern in the receiver. According to Shannon’s theory, informa= tion=20 “removes uncertainty” so that the more bits transmitted, the m= ore accurate is=20 the data pattern reconstruction in the receiver. Once again, removing bits= from=20 the transmission through data compression must “increase uncertainty= ” and lead=20 to data distortions or loss of resolution.=20

The only recourse = open to=20 the design engineer is to hide the image distortions, caused by image=20 compression, from the human observer. In the MPEG-2 compression standard, = for=20 example, 8 by 8 pixel tiles are converted into an 8 by 8 pattern of spatia= l=20 harmonic brightness sine waves, using the cosine transforms. The most domi= nant=20 harmonics are then selected by quantization, while the lesser harmonics ar= e=20 discarded. This distorts the images with blurring, visible tile boundaries= and=20 introduced image artifacts, such as light or dark lines. The image distort= ions=20 will increase with the compression ratio until the image quality becomes=20 unacceptable to the user. Distorted images may be acceptable to some degre= e in=20 television but not for scientific or medical images. The MPEG-2 standard a= lso=20 reduces the frame rate by sending only a single compressed frame per secon= d. The=20 missing frames are simulated by the receiver using motion vectors and=20 interpolation. This produces jerky and blurry motion within the images. Hu= ffman=20 coding is used to further compress the transmission by using shorter codes= for=20 the most frequently transmitted codes. This causes error propagation in wh= ich=20 any transmission error can cause an image to break up into random noise un= til a=20 recovery code is detected.=20

=20

Figure 2. An Autos= ophy=20 communication=20

As shown in Fig. 2= , true=20 communication is a process of copying knowledge from one autosopher to ano= ther=20 autosopher. The ultimate purpose is to “create new knowledge” = in the receiver so=20 that “information entities” can multiply and spread themselves= through the=20 environment. Transporting data or video through communication networks is = a=20 special case involving a transmitter autosopher and a receiver autosopher.= =20 Information is communicated with address tokens, called “Tips”= , each of which=20 may represent any amount of data. All input data is treated as “Addr= esses”=20 which, in contrast to Shannon’s quantities, have “meaning̶= 1;. Both the transmitter=20 and the receiver require a “hyperspace knowledge library” or m= emory, the=20 construction of which is explained in the next chapters. There are six kno= wn=20 different hyperspace library models or “Omni Dimensional Networks= 221;, all of which=20 can be used for data compression or different learning modes.=20

A “storyR= 21; in a book is an=20 example of an “information entity” copying itself into a reade= r’s mind. The=20 story is not dependent on other physical elements: matter, forces or dimen= sions.=20 However, all information entities require a physical structure or vessel, = such=20 as paper and ink in a book. The story has no mass or weight and scrambling= the=20 characters in a book does neither increase nor decrease the weight of a bo= ok.=20 Once a story is printed in the book it neither consumes nor generates ener= gy.=20 The story does not occupy time or space being both everywhere and nowhere = in the=20 book. The story consists of symbols in a progression, such as characters o= r=20 sentences in written text. During the reading process the story will copy = itself=20 into the reader’s mind, in effect creating new knowledge in the read= er. Once the=20 story has copied itself into the reader’s mind it will combine with = other=20 stories already in the mind to influence his behavior towards a political= =20 ideology, a new religious believe or computer expertise. Information entit= ies=20 may reside in books, in physical structures or as genetic codes in living= =20 beings. They provide direction and order in the cosmos, ultimately restori= ng the=20 order lost by entropy. Only information entities evolve towards more compl= ex=20 entities. Physical or living creatures only serve as temporary=20 vessels.=20

The efficiency and= =20 compression ratio of an autosophy communication depend on the shared knowl= edge=20 in the transmitter and the receiver. The more shared knowledge, the more=20 efficient the communication and the higher the potential compression ratio= . Only=20 the amount of “shared” knowledge in the transmitter and receiv= er is relevant;=20 not the total amount of knowledge in the libraries. Humans, for example,=20 communicate in a different manner with children who have little knowledge,= than=20 they do when communicating with experts in a field where both have a lot o= f=20 shared knowledge. Communications between experts in a field is much more=20 efficient than communication between children. If no shared knowledge exis= t, for=20 example when communicating in an unknown foreign language, then communicat= ion is=20 very difficult. Communication will become progressively more efficient as = a=20 common set of symbols is learned.=20

A communication &#= 8220;Tip”, in=20 order to contain true “information”, must contain something no= t already known by=20 the receiver, which relates to something already known by the receiver. In= a=20 communication every tip transmission could generate a new “engram= 221; of knowledge=20 in the receiver. An “Engram” is a unit or quantum of knowledge= which, according=20 to autosophy, is equivalent to a location in omni dimensional hyperspace. = For=20 example, telling a person something he already knows does not contain=20 “information”. Telling a person something completely unrelated= to already=20 established knowledge is gibberish and again contains no “informatio= n”.=20 Gibberish may be caused by unknown symbols, such as words in a foreign lan= guage,=20 or concepts too advanced for the receiver. In a perfect communication with= out=20 repetitions or gibberish each word or sentence would create a new engram o= f=20 knowledge in the receiver as a minimal extension to that which is already = known=20 by the receiver. The learning process may be imagined like the growing of = “data=20 crystals” or “data trees” in an electronic memory, one n= ode or one branch at a=20 time.=20

=20

Figure 3. Comparin= g Shannon=20 and Autosophy information theories=20

As shown in Fig. 3= , Shannon=20 and Autosophy information theories and communications are based on entirel= y=20 different concepts and lead to entirely different technologies. In autosop= hy all=20 input data (such as text character or pixel brightness values) are treated= as=20 “Addresses” which have “meaning” rather than meani= ngless “quantities” in=20 Shannon’s theory. Shannon in effect translates quantities and symbol= s into=20 binary numbers which are easier to transmit. Addresses are learned in a=20 hyperspace knowledge store or library in which the more knowledge already= =20 stored, the higher the data compression. Unlike conventional data processi= ng,=20 arithmetic is seldom used in an autosopher. Transmission is with Tips or a= ddress=20 tokens, each of which may represent any amount of data, for example, any n= umber=20 of characters in text or any size portion of the image in television. In a= =20 Shannon communication the bit rate depends only on the data volume or hard= ware,=20 while the data content is irrelevant. In an Autosophy communication the bi= t rate=20 only depends on the data content, while the data volume or hardware parame= ters=20 become irrelevant. The purpose of a Shannon communication is to “Rem= ove=20 Uncertainty” in the receiver, while the purpose of an Autosophy comm= unication is=20 to “Create new knowledge” in the receiver.=20

Shannon data stora= ge is=20 linear and dependent on the bits and bytes in the data record. Autosophy d= ata=20 storage depends only on the meaning of the data, which is stored in Omni=20 Dimensional Hyperspace. The more data already stored in the hyperspace rec= ord,=20 the less additional storage is required to store additional data records. = One=20 cannot learn what one already knows. Shannon uses equations to express ban= dwidth=20 and information, while autosophy only uses algorithms. An Autosopher datab= ase is=20 educated very much like a human child, a process which requires no program= ming=20 or data processing. Shannon’s theories will ultimately lead to the p= rogrammed=20 data processing “computer”, while the new Autosophy informatio= n theory will=20 ultimately lead to a next generation of brain-like self-learning=20 Autosopher.=20

THE RULES OF=20 LEARNING=20

“Learning= 221; is defined as the=20 accumulation of “knowledge”, Since communication depends on pr= ior knowledge of=20 the data, the learning process must be defined mathematically to be implem= ented=20 in electronic autosopher. The learning process may be imagined like “= ;data=20 crystals” or “data trees” growing in an electronic memor= y, one engram or one=20 node at a time.= =20

1 All input data i= s treated=20 as “addresses” which define “meaning”. Addresses c= annot be successfully used in=20 arithmetic or data processing, so conventional computing is unsuitable in = an=20 autosopher. Written text for example consist of characters, words and sent= ences=20 which have meaning. Using arithmetic to understand written text will not b= e very=20 successful. Our own brain is an autosopher which extracts meaning from=20 environmental data while hardly ever using arithmetic.=20

2 Communication &#= 8220;Tips” may=20 represent increasing amounts of data depending on the amount of shared kno= wledge=20 in the transmitter and receiver. The total amount of knowledge is much les= s=20 important then the amount of shared knowledge. For example, two very intel= ligent=20 persons may not be able to communicate if they do not speak the same langu= age.=20 Communications between adults, which have more knowledge, are usually much= more=20 efficient then communication between children, which have less=20 knowledge.=20

3 Communications s= ymbols=20 (Tips) may form progressions in which the same rules may apply on many lev= els.=20 In written text for example:=20

serial bit sequenc= es will=20 form ASCII character codes in a fixed length 8 bit code;=20

serial sequences o= f=20 character codes terminated by a space character will form=20 words;=20

serial sequences o= f words=20 terminated by a full stop will form sentences;=20

serial sequences o= f=20 sentences terminated by a new line will form paragraphs;=20

while serial seque= nces of=20 paragraphs terminated by a new heading will form sections.=20

Smaller or lower l= evel=20 symbols assemble in the same manner into higher and higher level symbols. = Each=20 string of lower level symbols is terminated by a specific “end of se= quence” code=20 or by a fixed string length limit, to form a higher level symbol code. The= many=20 levels in the progression may be encoded using the same learning algorithm= s. All=20 symbol codes on all levels can be interleaved in the same storage device. = Since=20 symbol or tip codes are used in the communication, any tip code may repres= ent=20 any amount of data, from single ASCII character to whole sections. There a= re six=20 different kinds of leaning modes or Omni Dimensional Networks each of whic= h=20 forms a different kind of progression.=20

4 All symbols (Tip= s) can be=20 defined like points in omni dimensional hyperspace. “Learning”= can be treated=20 mathematically as the defining locations in a space with arbitrarily many= =20 dimensions. While this process is very complex it can be defined by simple= =20 algorithms. Following the steps in the algorithms will implement learning = in=20 electronic autosopher. Each of the six known learning modes defines locati= ons,=20 using a different algorithm, in a different hyperspace model.=20

5 An “engram= ” is the=20 smallest unit or quantum of knowledge equal to a defined location in omni= =20 dimensional hyperspace. The amount of knowledge stored in an autosopher is= equal=20 to the number of nodes in its library. Each node is unique in the library = and=20 can be defined only once. This is because one cannot learn what one alread= y=20 knows. Each engram may represent any amount of data, from single ASCII cha= racter=20 to whole sections in text, or any sized portion of an image in television.= The=20 “meaning” of a symbol is specified by locating its location in= omni dimensional=20 hyperspace.=20

6 A new engram of = knowledge=20 is created in an autosopher only if the input information is not already k= nown=20 and relates to already established knowledge so that it represents a minim= al=20 extension to what is already known. Communications which are already known= do=20 not contain information. Communications which do not relate to established= =20 knowledge are gibberish and again contain no information. The bandwidth of= an=20 autosophy communication is based on the novelty and movement in the data o= r=20 television images. Everything which is not new or which does not move cont= ains=20 no information and need not be communicated.=20

7 Different learni= ng modes=20 are based on different numbering systems. This may be used to discover=20 additional learning modes in the future. Each learning mode defines quanti= ties=20 or addresses in a different way. The six already known learning modes are:= =20 serial (text learning), parallel (images, television), associative (databa= se=20 associations), interrelational (language understanding), logical (logical= =20 reasoning) and primary (unstructured access). Only the most primitive, the= =20 “serial” network, will be explained in the next chapter.=20

SERIAL SELF-LEA= RNING=20 AUTOSOPHY TREE NETWORKS=20

Self-learning Auto= sophy tree=20 networks may grow from text, speech or images in an electronic memory, lik= e=20 “data trees” or “data crystals”. The process is gu= ided by simple algorithms and=20 does not require programming or outside supervision. The result is true=20 mathematical “learning” that is a mathematical analog to the l= earning in our own=20 brain.=20

=20

Figure 4. A serial= Autosophy=20 tree network=20

Fig. 4 illustrates= the=20 serial network used in text learning. The same network may also be adapted= with=20 slight variations for speech or images. In written text the input data con= sist=20 of ASCII characters and the “end of sequence” code is a “= ;space” character. In=20 speech the input data consists of analog samples of the voice and the R= 20;end of=20 sequence” code is a zero crossing of the analog wave. In images the = data consist=20 of a spiral scanning sequences and the “end of sequence” is a = fixed length=20 string limit.=20

The tree network c= onsist of=20 identical nodes where each node represents an engram of knowledge. A GATE= =20 represents the new information which in text is an ASCII character. A back= ward=20 POINTER points to the previous node in the tree which represents the refer= ence=20 to the already established knowledge. The ADDRESS represents an engram as = a=20 point in omni dimensional hyperspace. Each ADDRESS is unique and can be de= fined=20 only once because one cannot learn what one already knows. A combination o= f GATE=20 - POINTER is called a MATRIX. What is finally stored in the memory is true= =20 mathematical knowledge. The network starts growing from a pre-defined SEED= node.=20 Several network levels in a progression may be interleaved in the same mem= ory=20 device by reserving a separate SEED node for each level. The first 256 mem= ory=20 addresses may also be reserved as SEEDS for the first string character so = that=20 each ASCII character forms its own sub-tree.=20

SERIAL NETWORK = GENERATION=20 ROUTINE=20

MATRIX: [ GATE ] P= OINTER=20 ]=20

Start: Set POINTER= =3D SEED =3D=20 0.=20

Loop: Move the nex= t input=20 character into the GATE.=20

If the GATE is an = “end of=20 sequence” (an ASCII space) then use the POINTER as output code; Goto= =20 Start.=20

Else search the li= brary=20 memory for a matching MATRIX.=20

If a matching MATR= IX is=20 found then move the ADDRESS where it was found to the POINTER; Goto=20 Loop.=20

Else, if a matchin= g MATRIX=20 is not found, then store the present MATRIX into a next empty memory=20 ADDRESS;=20

Move the memory AD= DRESS=20 where it was stored to the POINTER; Goto Loop.=20

SERIAL NETWORK = RETRIEVAL=20 ROUTINE=20

MATRIX: [ GATE ] P= OINTER=20 ]=20

Start: Move the in= put code=20 to the POINTER.= =20

Loop: Use the POIN= TER as a=20 memory ADDRESS to fetch a new MATRIX from the library.=20

Store, push the ne= w GATE=20 into a First-In-Last-Out (FILO) stack.=20

If the new POINTER= =3D SEED =3D=20 0 then pull the output data from the FILO stack; Goto Start.= =20

Else Goto=20 Loop.=20

Each operation is = started by=20 combining the first input GATE character with the SEED POINTER to obtain a= first=20 MATRIX. The memory is then searched to find a matching MATRIX already stor= ed in=20 the library. If a matching MATRIX is found then the memory ADDRESS in whic= h it=20 was found is used as the next POINTER in the MATRIX. If a matching MATRIX = is not=20 found, then the current MATRIX is stored into a next empty memory location= and=20 the ADDRESS into which it was stored is used as the new POINTER in the MAT= RIX.=20 The same steps are repeated for each input character in the string or text= word.=20 Eventually an “end of sequence” terminates the operation. This= “end of sequence”=20 code may be an ASCII “space” character in text or a maximum st= ring length limit.=20 The last node ADDRESS at the “Tip” of the branch defines the e= ntire input string=20 which is used in compressed communications. The receiver, which has an ide= ntical=20 tree library, retrieves the string in reverse order by following the POINT= ER=20 trail back to the SEED node. In a progression of symbols the code outputs = from a=20 lower level network are fed as input codes to the next higher level networ= k. All=20 networks in the progression may use the same basic algorithm and share the= same=20 memory device using a separate SEED node. In data retrieval each “Ti= p” address=20 is followed backward to the SEED node of that network level in the progres= sion.=20 The output codes retrieved from a higher level network are fed as input co= des to=20 the next lower level network until the final data is retrieved as output f= rom=20 the lowest level network. The “end of sequence” codes for each= level must be=20 added to the output codes from each level during retrieval.<= FONT=20 face=3D"Times New Roman,Times">=20

THE EVOLUTION O= F LOSSLESS=20 DATA COMPRESSION TECHNIQUES=20

The history of dat= a=20 compression is not a linear evolution but a twisted path followed by sever= al=20 independent researchers. David Huffman in 1952 invented a first lossless d= ata=20 compression method (Patent 3,694,813) based on the relative character=20 frequencies in a communication. Klaus Holtz in 1974 revealed the Huffman c= ode as=20 only a primitive example of an autosophy communication, opening the way to= more=20 advanced codes. Tree network data compression based on the new autosophy=20 information theory was first described in a patent application in 1975 by = Klaus=20 Holtz (Patent 4,366,551). A similar trail of research was started by Jacob= Ziv=20 and Abraham Lempel in 1977 and published just one month after the first Ho= ltz=20 publication. The Ziv-Lempel paper, “A Universal Algorithm for Sequen= tial Data=20 Compression”, describes an algorithm which is now known as the LZ-1 = (Lempel Ziv=20 #1) or LZ-77 (published in 1977) code. While Ziv and Lempel sought no pate= nt=20 protection, Douglas Whiting later received important improvement patents (= Patent=20 5,003,307 and 5,016,009). In 1978 Ziv and Lempel published an improved alg= orithm=20 “Compression of Individual Sequences via Variable Rate Coding”= , which is now=20 known as the LZ-2 (Lempel Ziv #2) or the LZ-78 (published in 1978) code. T= he=20 LZ-78 code builds a self-growing library by adding, for each transmission,= a new=20 string to the library that consists of a previous string appended by a new= =20 character. This code is rarely implemented because of its ever increasing = string=20 length in the library. In 1981 Willard Eastman together with Ziv and Lempe= l=20 (Patent 4,464,650) built a tree network using address multiplication. In 1= 984=20 Terry Welch (Patent 4,558,302) improved on the Eastman code in a paper = 220;A=20 Technique for high-performance Data Compression”. The LZW (Lempel Zi= v Welch)=20 code in effect re-discovered a primitive version of the Holtz serial tree= =20 network ten years later. Instead of terminating a string by an “end = of sequence”=20 code the LZW code terminates a string by a “first not found nodeR= 21;. In 1986=20 Victor Miller (Patent 4,814,746) improved the code by adding “delaye= d=20 innovation” and a recycling library. In 1990 British Telecom establi= shed the=20 international V.42bis standard for modems. The standard is a combination o= f many=20 inventions and improvements. Data compression is based on the serial tree= =20 networks discovered by Holtz (Patent 4,366,551) and modified by Welch (Pat= ent=20 4,558,303). Delayed innovation and a limited recycling library is used acc= ording=20 to Miller (Patent 4,814,746). A fast tree searching method, originally des= cribed=20 by Edward Sussenguth in 1963, is used to accelerate library searching. Mor= e=20 advanced data compression methods, including image and video compression, = are=20 now being developed by many researchers.=20

The following will= explain=20 the logical evolution of lossless data compression applications, from the = most=20 primitive to advanced future implementations, but not necessarily in=20 chronological order. It will point out logical flaws in the older codes th= at may=20 be used to improve the compression efficiency.=20

=20

Figure 5. The Huff= man=20 code=20

The Huffman code (= Fig. 5)=20 first creates a table showing the relative frequency of each character in = the=20 transmission. The least often used character frequencies are then combined= to=20 form a new table entry, which is again combined with the lowest character= =20 frequencies until each character has an assigned code. The result is a tab= le in=20 which the most often used characters have a shorter code than less often u= sed=20 characters. In a transmission system the ASCII input characters are used a= s an=20 address to fetch variable bit length codes from the table. The receiver ha= s a=20 much larger table in which most locations are empty while some locations c= ontain=20 an output ASCII character. The table is addressed in a tree fashion, using= the=20 transmitted bits as an address. The first input bit is used as the most=20 significant address bit, with the remaining address all at zero. If an ASC= II=20 character is stored in the addressed table location then the character is = used=20 as output data and the routine is restarted with the next transmitted bit.= If=20 the addressed table location is empty, then the next transmitted bit is us= ed as=20 the next lower significant address bit and the table is accessed again to = see=20 whether an ASCII character is stored there. Each input bit sequence will=20 eventually lead to a stored ASCII character in the list.=20

Huffman and Shanno= n argued=20 that rare or surprising communications contain more “informationR= 21; than common or=20 expected communications. The Huffman code is a typical but primitive autos= ophy=20 communication. It uses the input ASCII characters and the transmitted bits= as=20 “addresses” rather than binary “quantities” as in = a Shannon communication. Prior=20 “knowledge” of the data is required in the form of character f= requency tables.=20 If the frequency tables are wrong, because of changing data types, then &#= 8220;data=20 expansion” may occur producing larger output files then the original= input=20 files. A single bit error in the transmission will scramble all future=20 transmissions, an effect called “error propagation”. Because t= he relative=20 character frequency statistics represent only very scant “knowledge&= #8221; of the=20 data, the compression ratios are modest.=20

=20

Figure 6. The Ziv = Lempel 77=20 (LZ-77) code=20

The LZ-77 code (Fi= g. 6)=20 requires a library, implemented as a character shift register, in both the= =20 transmitter and in the receiver. The input data to be transmitted is shift= ed=20 into the transmitter’s library, while the retrieved output data is s= hifted into=20 the receiver’s library. The two libraries must remain identical duri= ng the=20 entire transmission. In a compressed communication the input strings are=20 compared with the library to find the “longest matching string”= ; in the library.=20 This longest matching string is identified with a START code pointing to t= he=20 first matching character in the string. A string LENGTH code is added to c= ount=20 the number of matching characters in the string. For example, the word ROS= E may=20 still be in the library from a previous transmission. Transmitting the wor= d ROSE=20 would require a string START code pointing at the first character R and a = string=20 LENGTH code 4 counting the number of matching characters. The input word R= OSE is=20 then shifted into the transmitters library. The receiver uses the transmit= ted=20 START - LENGTH code to retrieve the word ROSE from its own library and=20 subsequently shift the output word ROSE into its own library so that both= =20 libraries remain identical. A shift register library of 4k characters requ= ires a=20 12 bit string START code. Limiting the string length to 16 characters woul= d=20 require a 4 bit LENGTH code. A 16 bit START - LENGTH code transmission cou= ld=20 represent up to 16 text characters or a maximum of 8:1 compression compare= d to=20 normal 8 bit per ASCII character transmissions. The normal compression rat= io is=20 less than 2:1.=20

Ziv and Lempel bas= ed data=20 compression on “complexity” and previously transmitted “= data”, instead of=20 previous “knowledge” as in the autosophy theories. This theore= tical flaw may=20 explain the inferior performance of the algorithm. While previous “d= ata” may=20 contain “knowledge”, the two concepts are not identical. A shi= ft register=20 library is very inefficient; often storing identical strings in several co= pies.=20 If each string is stored only once in the library, then many more strings = can be=20 stored in a smaller library. This would increase the chances of finding ma= tching=20 strings and lead to higher data compression. Storing “knowledge̶= 1; in a tree=20 networks also eliminates the need for a string length code. Short data fil= es=20 cannot be compressed because few matching strings are likely to be found i= n a=20 mostly empty library. Pre-loading the library with a known data pattern be= fore=20 transmission would solve this problem. Pre-loading the library with a secr= et=20 pattern known only to authorized users results in a virtually unbreakable= =20 “codebook” encryption method. If no matching strings are found= , then each=20 character must be transmitted as a 9 bit code leading to a slight possible= data=20 expansion. A single error in the transmission makes the two libraries diff= erent,=20 scrambling all subsequent transmissions through error propagation. In spit= e of=20 its inferior performance compared with newer algorithms, the LZ-77 code is= still=20 the most often used compression algorithm.=20

=20

Figure 7. The Ziv = Lempel=20 Welch (LZW) code= =20

The LZW code, show= n in Fig.=20 7, starts from an empty library where the first 256 nodes are reserved to = act as=20 SEED nodes for each possible 8 bit character combination. This in effect c= reates=20 a separate sub tree for each possible input character. The first input cha= racter=20 in the string is used as a POINTER code pointing to the first 256 memory=20 addresses. The next input character GATE is combined with the POINTER form= ing a=20 MATRIX. The library memory is searched to find a matching MATRIX. If the=20 matching MATRIX is found then the ADDRESS in which it was found is used as= =20 POINTER in the next MATRIX. If a matching MATRIX is not found, then the st= ring=20 is terminated and the present MATRIX is stored in a next empty memory ADDR= ESS.=20 The next input character is used as the first character of the next string= . Both=20 the GATE and the POINTER in the present MATRIX are transmitted in the outp= ut=20 code. The receiver stores the received code in a next empty memory locatio= n in=20 its own library and then retrieve the output data string by following the= =20 POINTER trail back to the first 256 nodes.=20

LZW ENCODING=20 ROUTINE=20

MATRIX: [ GATE ] P= OINTER=20 ]=20

Start: Reserve the= first 256=20 memory locations ( 0 to 255) for the first character SEED.=20

Loop: Clear the PO= INTER in=20 the MATRIX.=20

Loop1: Move the ne= xt input=20 character to the GATE.<= /FONT>=20

If the POINTER =3D= clear then=20 move the GATE to the POINTER; Goto Loop1=20

Else search the li= brary=20 memory to locate a matching MATRIX.=20

If a matching MATR= IX is=20 found then move the memory ADDRESS to the POINTER; Goto=20 Loop1.=20

Else store the MAT= RIX into a=20 next empty memory ADDRESS;=20

Use the present MA= TRIX as=20 the output transmission code; Goto Loop.=20

LZW RETRIEVAL=20 ROUTINE=20

MATRIX: [ GATE ] P= OINTER=20 ]=20

Start: Reserve the= first 256=20 memory locations ( 0 to 255) as final SEED nodes.=20

Loop: Move the rec= eived=20 input code into the MATRIX.=20

Store the present = MATRIX=20 into a next empty memory ADDRESS.=20

Push the GATE into= a=20 First-In-Last-Out (FILO) stack.=20

Loop1: Use the POI= NTER as a=20 memory ADDRESS to fetch a new MATRIX from the library.=20

Push the new GATE = into the=20 FILO stack.=20

If the new POINTER= is less=20 than 256 then push the new POINTER into the FILO stack;=20

Retrieve, pull the= output=20 data from the FILO stack; Goto Loop.=20

Else Goto=20 Loop1.=20

The LZW code is a = serial=20 autosophy tree network in which each string is terminated by the first = 220;not=20 found node” instead of an “end of sequence” code. String= s or text words are=20 chopped into random fragments leading to a disorganized and inefficient &#= 8220;string=20 chop suey” library. Each transmission creates a new node in both lib= raries. The=20 libraries will keep on growing during a transmission until they becomes=20 unmanageable. A solution is to clear the library at a preset limit and gro= w a=20 new library. Short or random bit files cannot be compressed because few ma= tching=20 strings are found in a mostly empty or random library. A pre-loaded librar= y may=20 be used for data encryption. A single transmission error can scramble the= =20 remaining transmissions through error propagation. Overall the LZW code pe= rforms=20 only slightly better than the LZ-77 code. The original code had to be impr= oved=20 before it became commercially viable.=20

=20

Figure 8. The V.42= bis=20 compression standard using delayed innovation=20

The international = V.42bis=20 compression standard (Fig. 8) improves the LZW code through “delayed= innovation”=20 and a limited recycling library memory. Only the POINTER code is transmitt= ed,=20 while the next or un-matched character is used to start the next string. T= he=20 transmitter stores the not found node in a next empty memory ADDRESS and t= hen=20 transmit the POINTER only. The receiver retrieves the string from the POIN= TER=20 code and remembers the POINTER in a buffer. The next POINTER transmission= =20 contains as its first character the missing GATE, which is combined with t= he=20 POINTER in the buffer to create a new node in the receiver library. A rare= =20 problem exists because the transmitter adds a library node one transmissio= n=20 before the receiver can create the same node. If the transmitter uses the = new=20 node in encoding the next string, then the algorithm can crash because the= =20 receiver does not yet has that node. The V.42bis standard has precautionar= y=20 steps to avoid that problem. Instead of an endlessly growing library, the= =20 V.42bis standard has a fixed length library in which the “least rece= ntly used”=20 node address is cleared and recycled to store the next node. A few address= es=20 above the first 256 nodes are used as special command messages rather than= data=20 codes. The ETM (256) code switches the compression engine off. The compres= sion=20 efficiency is continuously monitored by an embedded microprocessor. If the= =20 compression ratio becomes too low, then the compression engine is switched= off=20 and replaced by clear 8 bit character transmissions. The compression engin= e may=20 be switched back on later when the data becomes compressible again. This a= voids=20 any possible data expansion. The FLUSH (257) code is used to clear the ent= ire=20 library to allow a new library to grow from the following transmissions. T= he=20 STEPUP (258) command increases the POINTER bit length by one bit. Starting= with=20 a 9 bit POINTER code (node addresses 259 to 511), the bit length is increa= sed=20 whenever the library doubles at 512, 1k, 2k and 4k. The V.42bis standard s= tuffs=20 either POINTER codes or 8 bit character codes into packets for transmissio= ns.=20 The packets contain error checking codes and defective packets are=20 re-transmitted until they are received without errors. This avoids the pro= blem=20 of error propagation.=20

=20

Figure 9. Sussengu= th tree=20 searching in V.42bis=20

The V.42bis standa= rd uses a=20 forward pointing tree searching algorithm (shown in Fig.9) to increase enc= oding=20 speed. Instead of having to search the entire library memory, this algorit= hm=20 limits the search to a maximum of 256 steps. Both the Sussenguth and the L= ZW=20 trees are combined in the same node. While the Sussenguth tree accelerates= =20 encoding speed the LZW tree accelerates retrieval speed. The GATE field co= ntains=20 the character which is shared by both trees. A DEPENDENT forward pointer p= oints=20 to the next node ADDRESS if the input character matches the GATE code. If = the=20 input character does not match the GATE code, then the SIBLING pointer is= =20 followed to the next node ADDRESS. The algorithm would follow the SIBLING= =20 pointer trail until a matching GATE is found. If no matching GATE is found= at=20 the end of the trail, indicated by an empty SIBLING pointer, then a new no= de is=20 created at a next empty or recycled memory ADDRESS. The previous node SIBL= ING=20 pointer is loaded with the new node ADDRESS to cut and relink the=20 trail.=20

The V.42bis standa= rd is a=20 state of the art algorithm, implemented in virtually all new modems, and w= hich=20 can compress most data to some extend. It avoids most data expansion and=20 eliminates the danger of error propagation by re-transmitting defective pa= ckets.=20 The algorithm is usually implemented by embedded microprocessors. The libr= ary=20 size, packet size and baud rate are negotiated by the modems before each=20 transmission. The Sussenguth search tree allows data transmission speeds=20 sufficient for telephone modems and ISDN though not much beyond that. Trul= y high=20 speed transmissions via local or remote area networks (including fiber opt= ic=20 transmissions and real time file compression in a computer) are not possib= le=20 with today’s microprocessors. Encrypted data cannot be compressed an= d no=20 provisions for data encryption are provided in the standard. Because of th= e=20 inefficient “string chop suey” library in the LZW code the com= pression=20 performance remains mediocre.=20

=20

Figure 10. Fixed t= ree=20 library compression=20

A fixed library sy= stem=20 (shown in Fig. 10) uses a fixed tree network to store a dictionary of word= s in a=20 known language. The library may be generated in a laboratory for use in ge= neric=20 open communications. This library then becomes a part of a communications= =20 software package, such as an Internet communications package, or the libra= ry can=20 be frozen in a Read Only version in a hardware compression chip set. A sec= ret=20 library may also be grown by the user for encrypted communications. A libr= ary=20 copy is then provided to authorized users only in the form of floppy discs= ,=20 PCMCIA cards or as a library download via the Internet. An encryption libr= ary=20 provides virtually absolute communications security, via public networks s= uch as=20 the Internet or cellular telephones. It also provides an impenetrable barr= ier=20 against unauthorized access to compressed computer files.=20

A library is gener= ated by=20 the algorithm shown below, from old text files. A different library may be= grown=20 for each foreign language. The library is usually grown by a software pack= age in=20 a PC and requires about 30 Minutes for completion. Old text files are fed = to the=20 algorithm to generate a first pass tree network. Only text words are learn= ed=20 where all characters are converted to lower case ASCII. The output word co= des=20 (tips) from the tree network are fed to a bubble sorting list. If the word= code=20 is already stored in the list then it is swapped with the code one locatio= n=20 higher in the list, unless it is already at the top of the list. New word = codes=20 are added to the bottom of the list. In this way each text word is stored = only=20 once and the most often used words tend to migrate towards the top of the = list.=20 The word codes are finally retrieved as clear text words through the tree= =20 network starting from the top of the list and stored in a computer file. T= he=20 text words can then be edited manually to remove word endings (such as = 220;-ing”=20 from verbs or “-ation” from nouns) and to split up compound wo= rds. This edited=20 list is then used in a second pass to grow the final tree network library.= The=20 library may be topped off to 8k nodes by a word list provided in the softw= are=20 package.=20

LIBRARY GENERAT= ION=20 ROUTINE=20

MATRIX [ GATE ] PO= INTER=20 ]=20

Start: Pre-load lo= cations 0=20 to 255 with GATE =3D ASCII code, POINTER =3D 0.=20

Loop: Set POINTER = =3D SEED =3D=20 0; Set string length =3D 0.=20

Loop1: Move the ne= xt input=20 character into the GATE; Increment the string length count=20

If the GATE is not= =20 alphabetic (A to Z) then output the POINTER to the BUBBLE LIST; Goto=20 Loop.=20

Else convert the G= ATE to=20 lower case ASCII (set ASCII bit 6 =3D 1).=20

Search the memory = for a=20 matching MATRIX.= =20

If a matching MATR= IX is=20 found then move the memory ADDRESS to the POINTER; Goto=20 Loop1.=20

Else store the pre= sent=20 MATRIX in a next empty memory ADDRESS; Move ADDRESS to=20 POINTER.=20

If string length = =3D 16 then=20 output the POINTER to the BUBBLE LIST; Goto Loop.=20

If POINTER =3D 819= 1 (8k memory=20 full) then end of learning; Goto Exit.=20

Else Goto=20 Loop1.=20

FIXED LIBRARY E= NCODING=20 ROUTINE=20

MATRIX [ GATE (8 b= it) ]=20 POINTER (13 bit) ] SPACE (1 bit) ]=20

Start: Clear the P= OINTER and=20 the SPACE bit.=20

Loop: Move the nex= t input=20 character into the GATE.=20

If the GATE is not= =20 alphabetic (A to Z) or an ASCII “space” then:=20

If POINTER =3D cle= ar then=20 output the GATE code (9 bit); Goto Start.=20

Else output the PO= INTER,=20 SPACE =3D 0; Then output the GATE (9 bit); Goto Start.=20

Else if GATE =3D A= SCII “space”=20 then output the POINTER, SPACE =3D 1; Goto Start.=20

Loop1: Search the = memory for=20 a matching MATRIX (use a CAM for fast search).=20

If a matching MATR= IX is=20 found then move the memory ADDRESS to the POINTER; Goto Loop.=20

Else output POINTE= R, SPACE =3D=20 0; Clear POINTER and SPACE; Goto Loop1.=20

FIXED LIBRARY R= ETRIEVAL=20 ROUTINE=20

MATRIX [ GATE (8 b= it) ]=20 POINTER (13 bit) ] SPACE (1 bit) ]=20

Start: Move the in= put code=20 to the POINTER, SPACE.<= /FONT>=20

If the POINTER is = a 9 bit=20 single character code then output the data character; Goto=20 Start.=20

If SPACE =3D 1 the= n push an=20 ASCII “space” into a First-In-Last-Out (FILO) stack.=20

Loop: Use the POIN= TER as a=20 memory ADDRESS to fetch a new MATRIX from the library.=20

Push the new GATE = into the=20 FILO stack.=20

If the new POINTER= =3D 0 then=20 retrieve, pull the output data from the FILO stack; Goto=20 Start.=20

Else Goto=20 Loop.=20

Test results show = that an 8k=20 node library is equivalent to about 130 Thousand characters of random text= (as=20 used in the LZ-77 code) and that it contains about 6000 of the most common= words=20 used in English communications. About 97% of the words used in a normal In= ternet=20 communication are contained in the list. Words later not found in the libr= ary=20 are automatically chopped into known fragments for transmission. This may= =20 require several transmission codes for an unknown word, but such word=20 fragmentation will not significantly reduce the compression because of the= =20 relative rarity of such words.=20

A text compression= and=20 encryption system can be implemented on a single silicon chip to improve=20 communications in modems or to compress disc files in a PC. A universal ch= ip set=20 can be used for open communications while a downloadable library can be us= ed for=20 encrypted communications. A Content Addressable Memory (CAM) can be used t= o=20 increase encoding and retrieval speed to more than 50 Million characters p= er=20 second, fast enough for virtually any application. A fixed library system = does=20 not suffer from error propagation and transmission errors will only cause = local=20 data errors. Data compression is active immediately, providing high data=20 compression for any length file transmission.=20

A communication pr= otocol may=20 use a 9 bit code for incompressible single characters and an 18 bit Tip co= de=20 which usually represents a whole text word including the space character. = A=20 CONTINUE bit separates single character codes from Tip codes. This bit may= be a=20 modified parity bit in storage devices or a modified stop bit in RS-232=20 protocols. For totally random bit files, the maximum data expansion is abo= ut 12%=20 while the average compression ratio for English text is between 2.8 : 1 an= d 3.3=20 : 1. A “space” (S) bit is used to indicate whole word codes or= word fragments.=20 The system can be improved through the use of a dynamic fragment learning= =20 library, which learns word fragments during a transmission for later encod= ing.=20 For normal text transmissions such additional hardware has been shown to o= nly=20 marginally improve the compression ratio. Two extra bits, the A and the F = bit,=20 are used to identify capital or upper case ASCII codes. This improves the= =20 library efficiency because only lower case letters are encoded in the tree= =20 library. If both bits are set then the Tip code represents a special messa= ge to=20 the receiver. This is used to insert error detection codes into the transm= ission=20 for automatic error checking in the receiver. It can also identify strings= of=20 identical characters or special messages to the receiver=20 program.=20

=20

Figure 11. Autosop= hy image=20 and video compression=20

Lossless autosophy= image and=20 video compression, shown in Fig. 11, is being developed for use in the Int= ernet,=20 video-conferencing and High Definition Television (HDTV). The hyperspace l= ibrary=20 contains a serial network which is adapted for image compression. The same= basic=20 scheme may be adapted for compressing static images or moving=20 television.=20

A first step is to= generate=20 a fixed tree library adapted for image compression. An imaginary phantom l= ibrary=20 contains the first two pixel brightness values or the SEED for each string= ..=20 This library contains all the possible brightness combinations of two 8 bi= t=20 pixel. Since the library only contains its own address no actual memory is= =20 required. A phantom library node is identified by the most significant add= ress=20 bit equal to zero. The second 64k library section contains the actual fixe= d tree=20 library where each node is identified by the most significant address bit = equal=20 to one. This library contains many thousands of pixel brightness strings w= hich=20 are pre-learned from normal input images, from a TV camera or from picture= s on a=20 CDROM. The 64k by 25 bit output library is finally made a part of a=20 communication software package or made available only to selected users fo= r=20 encrypted communications. A single chip hardware library, containing a Con= tent=20 Addressable Read Only Memory (CAROM), is used for real time television. Th= e=20 library may contain 8 bit per pixel absolute brightness values. The same l= ibrary=20 is shared for all colors (red green blue). A more efficient library may co= ntain=20 brightness differences, either the difference between the pixel and the ab= solute=20 center pixel brightness, or the difference between the pixel and its prece= ding=20 pixel. This highly compact library can be shared for encoding any pattern= =20 regardless of the overall brightness level in the images.=20

A superpixel or ti= p code=20 describes a small tile on the screen which is scanned in a fixed spiral pa= ttern.=20 The scanning sequence starts from a center pixel screen address. For stati= c=20 images the center pixel screen address for each tile is computed or taken = from=20 look up tables depending on the screen format. For moving television the c= enter=20 pixel screen address is a pixel whose brightness value has changed since t= he=20 previous input scan. To avoid any hardware dependency, i.e. the number of = rows=20 and columns on the screen, relative scanning is used. To identify a pixel = on the=20 right the screen address is incremented while for a pixel on the left the = screen=20 address is decremented. To identify a pixel below a fixed column constant = is=20 added to the screen address while for a pixel above a fixed column constan= t is=20 subtracted. If the scanning sequence hits a physical screen limit, either= =20 left-right or above-below, then the scanning sequence is terminated. This = allows=20 each television camera or monitor to have a different number of rows and c= olumns=20 while communicating in a common format. The column constant is known by ea= ch=20 camera or monitor and depending on screen size and resolution. Relative sc= anning=20 allows television cameras and monitors to evolve into larger and larger im= ages=20 while maintaining a universal and backward compatible communications=20 protocol.=20

An image buffer me= mory holds=20 the absolute brightness values of the input image to be encoded. Each of t= he=20 three colors (red green blue) requires its own image buffer. The input ima= ges=20 from the television camera are scanned into the image buffer for encoding = in the=20 transmitter. In the receiver the output images are scanned from the image = buffer=20 to a monitor at regular scanning rates. Since only the moving portions of = the=20 images are selectively transmitted the scanning rates in the transmitter a= nd in=20 the receiver need not be identical.=20

In television a ch= ange=20 buffer memory holds the screen addresses of the pixels that have changed s= ince=20 the last input scan. While a new image is scanned from the camera each pix= el=20 brightness is compared with the pixel brightness in the image buffer. If t= he=20 brightness has not changed, within a threshold limit, then the input pixel= is=20 ignored. If the pixel brightness has changed then the new pixel brightness= is=20 entered into the image buffer. The screen address of that changed pixel is= then=20 stored in the change buffer. After the input scan the change buffer would= =20 contain all the screen addresses of the changed pixels.=20

For static image=20 transmissions or storage an input image is first scanned into the image bu= ffer.=20 The image is divided into tiles by computing the center address for each t= ile.=20 Each tile is then encoded into strings, by spiral scanning around the cent= er=20 address. The strings are then encoded into hyperspace INDEX codes using a = fixed=20 tree library. Each string is terminated either by a fixed string length li= mit or=20 by a not found node. Not found strings are encoded by string fragments unt= il the=20 string limit is reached. The transmitted 17 bit INDEX codes could each rep= resent=20 between 2 and 24 pixels. The compression ratio would range from a 6% data= =20 expansion for totally random patterns to 11 : 1 compression for blank imag= es.=20 All three colors would use the same fixed library.=20

For moving televis= ion the=20 input images from the camera are compared pixel by pixel with the previous= image=20 in the image buffer to identify the locations of changed pixels in the cha= nge=20 buffer. Each address in the change buffer is then encoded into INDEX codes= , by=20 relative spiral scanning, using a fixed tree library. Since change in movi= ng=20 television is caused by large moving objects in the images, the changed pi= xels=20 usually form clusters around the leading and trailing edges. Many changed = pixels=20 in a cluster may be encoded into a single INDEX code by erasing the pixel= =20 addresses in the change buffer which are identified in the spiral scanning= =20 sequence. Each tip transmission contains a (17 bit) INDEX code and a (20 b= it)=20 LOCATION on the screen code which identifies the center pixel (the pixel t= hat=20 has changed since the last scan). A more efficient library may contain=20 brightness differences which are shared for all light or dark image sectio= ns. An=20 absolute pixel BRIGHTNESS (8 bit) is then added to the transmission code. = Each=20 transmission tip then represents at least 3 pixels.=20

Comparing the MPEG= -2=20 compression method with Autosophy compression reveals the following=20 differences:=20

1) MPEG-2 and othe= r Shannon=20 type transmissions require a fixed hardware standard for screen size and=20 resolution. Autosophy transmissions, in contrast, can be made totally=20 independent of the physical hardware. This allows television technology to= =20 freely evolve into larger images, with better resolution, while maintainin= g a=20 common backward compatible transmission standard.=20

2) In MPEG-2 the i= mage=20 quality is determined by the required compression ratio. The higher the=20 compression the worse the image quality until the images become unacceptab= le to=20 the user. Autosophy in contrast offers entirely lossless compression which= will=20 not distort the images. Because random background noise is interpreted as= =20 movement any increase in the noise level will increase the transmission. B= etter=20 image quality with less noise will actually reduce the=20 transmissions.=20

3) Shannon or MPEG= -2=20 transmission require a fixed bandwidth channel which is determined by the= =20 systems hardware. High speed fixed bandwidth channels are expensive in pac= ket=20 switching networks. In Autosophy television, in contrast, transmissions de= pend=20 only on novelty and movement in the images. Slow moving or familiar image = will=20 produce few transmissions while rapidly moving complex images will produce= more=20 transmission codes. Autosophy transmissions are ideally suitable for the n= ew=20 packet switching networks, such as the Internet or ATM.=20

4) Any transmissio= n error in=20 MPEG-2 will cause the images to break up into random noise producing very= =20 disturbing visual effects. In Autosophy television, in contrast, transmiss= ion=20 errors will cause only a temporary freezing of motion in small image porti= ons.=20 Re-transmitting defective packets will correct any visual=20 effects.=20

5) MPEG-2 can be e= ncrypted=20 but only with additional hardware. Autosophy provides a virtually unbreaka= ble=20 encryption, without extra cost, by making libraries available only to auth= orized=20 users.=20

6) MPEG-2 compress= ion=20 requires very high speed computing for the cosine transforms and motion=20 compensation. Autosophy compression does not require any computing. It is= =20 expected to be much less expensive to implement.=20

SUMMARY AND=20 CONCLUSIONS= =20

High speed digital= =20 communication networks, such as the Internet or ATM, are rapidly evolving = into a=20 world wide Information Superhighway. This technological revolution is driv= en=20 mostly by the hardware which provides more and more bandwidth at lower and= lower=20 cost. A similar revolution is occurring in the communications software whi= ch=20 provides better and better protocols and higher and higher data compressio= n.=20 Lossless data compression techniques are based on a new Autosophy informat= ion=20 theory which may eventually replace the classical Shannon information theo= ry.=20 Autosophy based data compression is already implemented in V.42bis modems = and in=20 computer disc file compression. It is rapidly spreading into virtually all= =20 communications networks. Autosophy data and video communications are espec= ially=20 suitable for the new packet communications networks. In addition to high d= ata=20 and video compression, Autosophy transmissions may also provide a virtuall= y=20 unbreakable encryption code for secure communications via public=20 networks.=20

The new Autosophy = theories=20 of “learning” are now evolving into a new generation of self-l= earning brain-like=20 and no-programming autosopher which may eventually replace the now dominan= t=20 programmed data processing computer. Such futuristic brain-like autosophy= =20 databases already exist in laboratory models that strikingly emulate learn= ing in=20 the human brain without programming or conventional data=20 processing.=20

Autosophy theories= may also=20 explain the functioning of self-assembling structures, such as crystals, l= iving=20 trees or societies. It adds “meaning” and “information&#= 8221; as a new element in the=20 cosmos. This research may have a very profound impact on other to natural= =20 sciences, such as physics, chemistry or biology.=20

REFERENCES=20

1 Klaus Holtz, = 220;Data=20 Compression in Network Design”. Tutorial at Design SuperCon’96= , Santa Clara CA,=20 Jan 1996=20

2 Klaus Holtz, = 220;Digital=20 Image and Video Compression for Packet Networks”. Tutorial at SuperC= on’96, Santa=20 Clara CA.=20

3 Klaus Holtz, = 220;Autosophy=20 Data Compression Accelerates and Encrypts Network Communications”. W= ESCON/95,=20 Session C3, Paper 4. Nov. 7, 1995=20

4 Klaus Holtz, = 220;Autosophy=20 Image Compression for Packet Network Television” WESCON/95, Session = C2,=20 1995=20

5 Klaus Holtz, = 220;”Packet=20 Video Transmission on the Information Superhighway Using Image Content Dep= endent=20 Autosophy Video Compression”. IS&T’s 48th Annual Conferenc= e, Washington DC,=20 May 1995=20

6 Klaus Holtz, = 220;Fast Data=20 Compression Speeds Data Communications”. Design SuperCon ‘95, = Santa Clara CA,=20 1995=20

7 Klaus Holtz, = 220;Lossless=20 data-compression techniques advance beyond Shannon’s limits”. = Personal=20 Engineering Magazine, Dec. 1994=20

8 WESCON/94, Sessi= on W23,=20 Advanced Information Management, Sept. 29, 1994, 5 papers, Anaheim=20 CA.=20

9 WESCON/94, Sessi= on W9,=20 Digital Video Compression. Sept. 28, 1994, Anaheim CA.=20

10 Colin Johnson, = “Data know=20 thyself” OEM Magazine, May 1994, Page 94=20

11 Klaus Holtz, &#= 8220;Hyperspace=20 storage compression for Multimedia systems”, IS&T / SPIE Electro= nic Imaging=20 Science and Technology, Paper 2188-40, Feb. 8, 1994, San Jose=20 CA=20

12 WESCON/93, Sess= ion 7,=20 Applications for lossless data compression. Sept. 28, 1993, San Francisco,= CA. 4=20 papers.=20

13 Klaus Holtz, &#= 8220;Autosophy=20 Networks yield Self-learning Robot Vision”. WESCON/93, Session S2, P= aper 5, San=20 Francisco, CA, Sept. 28, 1993=20

14 Klaus Holtz,=20 “Self-aligning and compressed autosophy video databases”. SPIE= Vol. 1908, 1993=20 San Jose, CA, Storage and Retrieval for Image and Video=20 Databases.=20

15 Klaus Holtz, &#= 8220;Lossless=20 Image Compression with Autosophy Networks”. SPIE Vol. 1903, 1993 San= Jose, CA,=20 Image and Video Processing=20

16 Klaus Holtz, &#= 8220;HDTV and=20 Multimedia Image Compression with Autosophy Networks” WESCON/92, Nov= . 1992,=20 Anaheim, CA, page 414=20

17 Klaus Holtz, &#= 8220;Autosophy=20 image compression and vision for aerospace sensing”. SPIE-92, Vol. 1= 700-39,=20 Orlando, FL, April. 24, 1992=20

18 Colin Johnson, = “Storage=20 technique can learn”. Electronic Engineering Times,=20 Jan.6,1992=20

19 NORTHCON-91, Se= ssion D4,=20 Learning Networks: An Alternative to Data Processing. Portland OR,=20 Oct.1991=20

20 Klaus Holtz, E,= Holtz,=20 “True Information Television (TITV) breaks Shannon Bandwidth Barrier= ”. IEEE=20 Transaction on Consumer Electronics, May 1990, Volume 36, Number=20 2=20

21 Klaus Holtz, &#= 8220;Text=20 Compression and Encryption with self-learning Networks”. IEEE GLOBEC= OM-85 New=20 Orleans=20

22 Klaus Holtz. &#= 8220;Build a=20 self-learning no programming Computer with your Microprocessor”. Dr.= Dobbs=20 Journal, Number 33, March 1979, Vol.4=20

23 Klaus Holtz, E.= Langheld,=20 “Der selbstlernende und programmier-freie Assoziationscomputer”= . ELEKTRONIK=20 magazin, Dec. 1978, Volume 14 and 15,Franzis Verlag Abt. Zeitschriften Ver= trieb,=20 Postfach 37 01 20, 8000 Munich 37, Germany=20

24 Klaus Holtz, &#= 8220;Here comes=20 the brain-like self-learning no-programming computer of the future”.= The First=20 West Coast Computer Faire 1977, Faire, Box 1597, Palo Alto, CA=20 94302=20

25 D. A. Huffman, = “A Method=20 for the Construction of Minimum Redundancy Codes” Proceedings of the= I.R.E. pp=20 1098-1101, Sept. 1952=20

26 C.E. Shannon, &= #8220;A=20 mathematical Theory of Communications”, Bell Telephone B-1598, Vol.2= 7, July and=20 Oct. 1948=20

27 J. Ziv and A. L= empel, “A=20 Universal Algorithm for Sequential Data Compression”, IEEE Informati= on Theory,=20 IT-23, May 1977= =20

28 J, Ziv and A. L= empel,=20 “Compression of Individual Sequences via Variable-Rate Coding”= IEEE Information=20 Theory, IT-24, Sept. 1978=20

29 T. A. Welch, &#= 8220;A Technique=20 for High Performance Data Compression” IEEE Computer, June 1984 (Pat= ent=20 4,558,303)=20

30 Y. S. Miller, M= . N.=20 Wegman. “Variation on a Theme by Ziv and Lempel”IBM Papers, Co= mbinatorial=20 Algorithms on Words, 1985=20

31 Edward H. Susse= nguth,=20 “Use of Tree Structures for Processing Files”Communications of= the ACM, Volume=20 6, Number 5, May 1963=20

Patents: 3,694,813= .=20 3,914,747. 3,976,844. 4,021,782. 4,038,652. 4,054,951.
<= FONT=20 face=3D"Times New Roman,Times">4,152,582. 4,366,551. 4,412= ,306.=20 4,436,422. 4,464,650. 4,558,302. 4,814,746.
4,870,415. 4,992,868. 5,003= ,307.=20 5,016,009. 5,051,745. 5,113,505.=20

------=_NextPart_000_0000_01C71122.93BDACB0 content-type: image/gif content-transfer-encoding: base64 content-location: file:///C:/Documents%20and%20Settings/Klaus%20Holtz/My%20Documents/My%20Web%20Sites/mysite3/THEORY/info1.gif R0lGODdhFwLoAPcAAAAAAIAAAACAAICAAAAAgIAAgACAgMDAwMDcwKbK8IAAAACAAICAAAAAgIAA gACAgICAgICA/4AgYP//wKDg4GAAgP+AgACAwMDA/wDP/2n//+D/4N2cs7OP7ipv+T+4zUiENpWM QY5eQqBiemJPrB0vvihmdgBFAEU+AWooE4U5akoyhQQEBAgICAwMDBERERYWFhwcHCIiIikpKTAw MF9fX1VVVU1NTUJCQjk5OQAHAA0AALeZgYSZtL29kH9/YGBgfwAOABsAACgAAAgJKwAdADkAAAAA mwAlAEkAABEROwAvAF0AABcXRQA6AEkRERwcUwAW/ysA/yEhbFkUFABRAEcaahkyZwBhAAAx/2EA /1MgexZDZy4u4iZZFlFGBGguSQdSj2oYuJAjFQBT/6MA/2pKEnUzbEpBmjdlC6QsFYMfsU4s/yBR tghkkm9WC1lDrTZyErAzFwChAHdfH4lHcbBDHLctfQCGlXpuIyafAHOpAQAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAADKAKxbASAdwpRScCSqTAqUiTZue0R1kP8AqABx/98A/1aRSjRI+MwyguRBcGjK ATa8QgCa/5Yit4V9MyW3jDZa7Vz/AP9IACKbokLPTcJYUiDTlaUk4HNWtampANBvPGefWInPC/+s AKcu/uJZf0zcZ/8Y/zp9/7HQGMf/AP/iAN+aPVaBn8ZDuq9xiziiydFTzv+aZUbK2/9N/8jpakze 4P+Y/9/AgunspfX2zf/Q/7GsWmORriJMZY1OP1BwcND////n/2lpaXd3d4aGhpaWlp2dnaSkpLKy ssvLy9fX193d3ePj4+rq6vHx8fj4+LLBZoC/eMbw8LKk//+z/9GOo8PcN6CeVHaucHiewYNkv6SD 09E/Mv99AER4IyRfYA4OLL4AAP8fADE5ANmFPgJ3hbDYgVYhHQAAMIjIs6B5AMDAwOpwgVHxaf// gJF0zf98/6L////78KCgpICAgP8AAAD/AP//AAAA//8A/wD//////ywAAAAAFwLoAAAI/gD/CRxI sKDBgwgTKlzIsKHDhxAjSpxIsaLFixgzatzIsaPHjyBDihxJsqTJkyhTqlzJsqXLlzBjypxJs6bN mzhz6tzJs6fPn0CDCh1KtKjRo0iTKl3KtKnTp1CjSp1KtarVnACy/gNAMKvWgV4Lht2qdSzYslwF mg2b9qrbt3DjZmz7dWvXuwbb2lWrlyxYvHv9yh1MuLDhvmn7BlYceLFjunT/SjZMubLlp4wbq50M +PHfyIm7Zr5MurRpnpkVR84LmCtoya7zjj5Nu7btlKlZb9a9e29sx3Zzd75NvLjxi4h7T2asd/Xv 1YIbq559vLr169Cldw69fDN37Xw5/ou+Tr489rpn0Xs1S7bu2MRoobtGf5C6+fv487ukr7+///8k 2QfggAQW2BB/Biao4IJnMejggwYKCOGEFN6GYIUYZkibhBp26OFbF34o4ohWhUjiiQ+up+KKLLbo 4oswxijjjDTSmNCKYrHH1nsqaobijzQxV9+QvA2nnHhHJunjkkIWieSSnGXX3HaQAWllTU2KRaSW TiqZpZFMbgnml09GmaSU45V55Zq4ifkkmV66GWeXYdIJJ5TKfdedd7uZyOafItUo6KCEFmrooGq2 516RoenpJ6CQbnTnpHLWyeWlY1ZKKW+zOcqob5ZGKqpHm2L6pqao2pmqqXPiCZyP/lM2mueotHZU aqaq5spqqKfq2ppCaCI5ZXhK1mqsQ4cmq+yyzLqYXohrQYbgsME9euy1vva6663a4rrtqr8Wi+24 M3HbqrfodqvuudpmR+675YK7Lq/s1kvvvVm6OxGH8PZbX7MAByzwjM/qC5G1/iacrb3m4isvw6vK VxG/Clfs8MIXf4txw/nuKRHFFivM8cMZpwvxxh6DjHDICQ/s8ssvFyzzwSzXfCPJI6Oss8Y8Syyu bDYHvbPJJc+bM8/hviohyEK/ezTRTxuN85Y+u5po0y3DrPXWyc6sqMFpYh101CcjLfXQ7aacKNNi X0t20WVDPTWnaovLdtvHvq33/txml121wSvj7TTfcqMd99k5Rqs4f3cLnjfhiBfeN9yW/h1l444b u7fhlG9+qeW9YZ655pAfbnrnfIMO6uhic+366zbytbi0rLftueS4R4646qLXTjrnt+t+esehe+x7 08EPXzrqdBfvPGU4OulefDpSL/tX0V+/I3ypZb/9eumFypb2zZVltX7JMz95+j1q7zX0RqqWNHiz Ugscd/bjRS3+fTaIaZXPA2DvbMO+5RVQV7y7TP5+479XPU87CzRV/h7oruQIazsXnN+ADrg+A3ow gMY7zHAYyCcHKm11ERxh/DB4pOncLGywWVvi2IOdD3YQeDZMmuqgB775yaqF/vpDoQpH2D4q6WhX 0SmTetYCrAjB7olQLKIUe3g+t9iviKCq4F1+SD8gsvCLJuwUEp0HthZisYY4TOMN1+itBFpmf59r oAVLyMAUimeCV7zj/ySoQc1M0D8czJ3y1OjGysDRh2baYyLPpEgwajFseKza6vwYoRwKT32CrFzd qliiKdKOiV/7Eu1CKZtR0lBaigvfZ+gjv9mVT34FCuQlZeklKnoycOMaYJsqqcZM0jJjhRxbUHS5 IUsOko2zpNomiQkpZgZqQb+Mpg2DebyxRfGaT/SaFKuJNWn2MpkIXCY3kWdMTIITmSBM5zht5k10 mtNv4jTkGanYHkWRz2r2/hnf9USDGOz5s5/zUaU+R9VOX5aTObZMKA0Lc8hZXa6BrCHh5zzVrfLl KJEAJOhBN/rNVlGToUGM40PXBi0twVJ8W/yMHB2q0Y6+86VH+yhhhnXS1/gmn8FBCJwAh0Gb5nQx zpxQQc9p0HCqM6hF2aYinwNDnTaxSzw1k0+5SCxaYfOqrtOmLRXIUqmWcIwHklNUHcrU/t3pT0M9 ZlF7Fk9DhhCCdATTQ5JTUzHRdJIClCub0gpTjp7wgSLs6kmBejWGSDKPXfxhsBZlVb+u9bG7a+th pvcsu66yd6esHj03iz3Z1dOsuCQRX0erKYVOb6HrrBhpXeo5maa2ZY4l/qps4XnU19Zste7koGtt Cy+s+hZmWgUlb1UbW7XOFpiSHa6/cAtZ49YWsMrtbXH7ylplPperOr3lafl5U+6RL6CeJaVnvQtQ V4YPldxNL496tDhIXohFAo3ejojC3OPqzbSoRCp9f/oplu4vWEJMbAzDk9FFOsd/iG0k/xyYYC41 lH9V0u8zq0vh3EYsuZMdTUP/arzBRgePKc0gdGPVvEbqEVYrLbEJBczJIE23vrTlsISFEhvhUPCt i9QjRS3a1C86CsTn8rCeijdkFqO4xzMm1YuXTEgMBza7mlXP5bi3UAuSkJ5VzfL7sCyYuvZ3iF6W qHQ2a6c/6uS3aBZY/nAZa5xDzvHNKwziXcWqGys71csm9mKYj+xfOkF3JzCmroWvm+SmPFiDAD5x Egd7ZRUzms+ilB4jQ/fotya40BwJtKbbiumlHDrHifaiYD9MapHW78Cg7qqoG1zpEUOysFhi8qBn 3UYnb8iTDYpyfF+pSqD2cD6NYhxl5fvZfRbsldMZpT17XVNcM7vYLZbJpisMNfy+L7pJQWqnJSXr 5gpah5s0LHvPaEaAihfbNLPVUKZN63Pu1rpAbiqbt227TO+X2vaV9buhqkK6qvRn6C5OmgfOrDWX kUihlmgd7xdw8rDb2zG1NZTza2r+1hG1DbdQt/Nd4X07eMD8lPJP/lGd8eM8nONr9Hicu8jw79Db z+LNnhETd+5j2xVHMkfa+GS+c2SJ++JLRGjQNRvtm5zcuZ2ztvvm6miRlpXlI6mxU28swX7iON4B BlukifW3IpsUT3nNcapBHtpY4xvp7dzhXIOe6+2+Z7zkDlBJv/xVi684z2Ke49TpXmQ4A2vu/+4j nidZdJsc/duRVSdppI5wMPLX7iTHO5/7B3Orh9RVGn68yfQ1+AIb5fC4Vfvi/z50v0O+sv86paKN jcTuoR7zCxEjvyfdGi6XHUsEz72hDF6eHZeSsU/3OuGXfuKzyt7lT6XITkW8p8x/fuNo9+thbz9T 6+o5rgBP+BD3/s5HkK++cf6uOO0PexTQQ1/G/zK59SmZRaoj0vF2jF/kka9Xn7sfPJdOcVLPb37F 05yArDRsp/V29ZRzxRYfu2ZzfQNs99RrsBZ7AVh6yJZefsGA5cd/GFhms/Nzo4NpL2ca/dduMcY0 wsZlfoRzcSciSfaBIJiBIihoBweBecZ+C1Z4ECJhLNiCZ4d40VdixBRBZJKDJScVuleEiEJ8c9Fv biJ8Q4hGLxiCtfYR88R8y9aE+AGFKBeFSgYrtvdxVniFLghx0wQS2rd9XwiGOxh6RmOCSahqegd1 Z9hmYZiF3tcwmSaAymZecehwc9iDX5dJeygyfciDSXQ6gZg1/kaYiDFyiIwIaI34iFAhhJA4iVtI iZb4fJeYieumiZwIFJLYiaAobqF4PN7jbAWYgKEEXgWIa5/EdKNYTRvWdT5GeZ8WdoX4Qq9Iiu83 fA1mRpZ2eVX0ibn4I7HofXQHRPTHYz2Gi8PoO1j0jHkygIzzceh1UdaYfM3ojOFiZ44kaXYHJWcF cNlYb+E2ZL3IfqXGQmZGfeNoMZ92f1hHP7HiHGPVjs5oigjIep9FdOHVXW/HRKgljPYIigI5kJxY kAaZkAq5kAzZkA75kBAZkRI5kfuiXa34epnlXakIbQnlgBRJjumUURtmYLQ4iw/4ke4oeMBIeTNo erARgyjJ/jJuJm91Zlndd2RPF5Otw1nClkFbt32wRFU6SU40OHVBaZNAWZOaN5TCpGr4p5QtKX54 hZBMmSKpNm/vF4/5d5JVSS7REnPQVoU8V2XbhZFs1pVomZZquXisqGx35GwZmUrnFV7wdYprWSvv SC/F+GX/1XxfxY53qSEPxnno+JRLCFiEKXaBCSiH1neFSWLch1fM6IuLiZdkZlYVN4XOR3bqVWaV 2VIheYJHdC/Fp2geVn+fiSLvKHLh9jPciI6tZoOpiSF5aUeJyX2JdpvpOJtXUooW6GvnRVn9OENl yXr8WC1UqVw8Epxy+T0Yx5vrxo/OST2m9E9sSBVbyWAl/gmdSMFq20l+2wkXyrhDb6gRiohVAnee 2LRy8cdhkxcXkEmeYDUxp5GcOGGfMRFv7QmeqImdcvZ6LBdU+AkTA2p2paGfZliFUFegJXFX9Vie bXigAlef7JmgP8mVmFF7FTWfykehxMGgLYGgq4dSHDoV2Ql/UuihhmWWGkkluwY+6wWj1NmKHAKi LKF3vQih8Qgi//SSQmdu6qaDsUeS6jKTkhlgDMd19iehM/R7/xeW+jicvamiuHiUFzSe3piUChZW Qopt5TWdS1idzxGBqzhuZqqHMOeUFBVi2JegsBmNlVKivyeNvEZlPVqmDWiAEWGjxOiFa8p8Lud5 wIeZ/kh6jvgUZ7YZQ0KWpbYHoXLKlykjn33mlOmGbjbmhidYh4G3cOlyqeGopm64VUtZnn7nqSu6 LyrZR/KILuDXcJcaoJQUqBG1qWZjjkTjeISXQp5XmKvGd6f6MduoJn+6WLhaqV4amVNIVplKq49U dRH4XkNqkpi6oKzikqpoZv05cRNoN6MZePqYgq5oqch6JpcpZi2HmIdZlJEJVXFJgd7zpMcJpsjJ oWxjq3fHkrHqhfjaoR+ip2OWrGQ4rtNKWKs6ciMqalIpjp86TFwKrKCqquGXNhaBbNB6bI5KpVtn rwGbrrBKVj4TfLUKbnG6LYBJoEu6p2OXm7SHURia/qUkqrEfykenuZTcVkrM9pvbo6hlSV7F2Y++ CVHiGLQ0Fq4Hg4fHGXOsRJwtSn1vGLGnwqf5KbOqol9Qu0sa16W8mndDUrUvgaMHC5P8yqRXK7YI 26w4hrUUW63XiRxUWhtcexJNG3KkurbYFUL7maJYW0xoK353O7Ydi7D49JwjC1LwKbjZCp8m5rVu eiCGqxRjKaY32yl1iY11y6P+KrTiKbe8RpfF2bhz6rnWASOT+UZyIbppWn1Y8SKAJCOneyPt07MD JZbP+LrgSrRWxLotW3OP+6/5RbF2illnhrv3cYR3w58EW7CTCofGeruCErbIK5K76J6iU6BHeJ/q /lm9DnuNj2SlV3e4russqTe5SpuCpvu5bFgoqJqqxsiLG4oc11sj50EwuYuvIEtnb2pkzsu8uFuv TqdEUElGxYqyqSu/vVe+j+qkiuq/+kpppYcRb2ueBoy5Nymqs/dmUQa8Awy++aHB6+qKDqrAKzl/ Ejy6nSS+B6ylWve/Sbqv9JnBtetwLzy/lbO9KjxW01u6Mcy/6svCGks8bOvCD+wToVW8wLiX8Hh5 N1y4y0u0iKWyo8aZDhy8QFKvtFtBJdi7muuzrRqzHzO7PInAulvFQJu9x5q3bUu6ARfEARJ1bcmK XXvG//HFYWl5+4iCwCajvzuXJ3wVKxishSqJ/mocdU2Kim/5ojkLXzHqdlE7diuXpAtWgys8jWYc pOgawHCLsd1YrX78yPRLdp/YeX+ofyOquLgayBN2ymEENIsstqDMplW1qGYYmzfadK5ssDdZtnDa wfDDxt1aLCUbxULayt5qyxa6VMs4y5mZle8ZtwCqZSPMx2tst5SrEqYcErd5okKZtf9DlqvcYXL7 n7f8mL7cuqUbzZWsmFY7euNnRCKntcW8vs9cid7cq87MzIq5o/nLthlpztaMaNr7xsFMy/d7i/b8 sPEszyeqneFMytVChcqnS4JqrvYWKPgoZaCLypXrnfS8stosVWDboGTZufkompFryMG2sLbbvcJ/ OXyUTEBt26gijbPwGtN5CqUAbWjm6a5rG7Gcird+W7kyfGvV7L3R6hnRCIS0urE/jcY2O04DJDHU pFhBXdQuPcm29dTNV7FFSX8+XdVkG11YTWRFbWdcLc9uC8e8hVk724VT1iedpdRerc7iKsijnNIT vdRulcb8/EcCMmND3aBofdUg3c6n28dc/NXKeaN0i8x4vctzjdiBTbiPLddxDdTK+b5QlJ6YDTvc 2dme/dmgHdqiPdqk7TgBAQA7 ------=_NextPart_000_0000_01C71122.93BDACB0 content-type: image/gif content-transfer-encoding: base64 content-location: file:///C:/Documents%20and%20Settings/Klaus%20Holtz/My%20Documents/My%20Web%20Sites/mysite3/THEORY/info2.gif R0lGODdhNQIJAfcAAAAAAIAAAACAAICAAAAAgIAAgACAgMDAwMDcwKbK8IAAAACAAICAAAAAgIAA gACAgICAgICA/4AgYP//wKDg4GAAgP+AgACAwMDA/wDP/2n//+D/4N2cs7OP7ipv+T+4zUiENpWM QY5eQqBiemJPrB0vvihmdgBFAEU+AWooE4U5akoyhQQEBAgICAwMDBERERYWFhwcHCIiIikpKTAw MF9fX1VVVU1NTUJCQjk5OQAHAA0AALeZgYSZtL29kH9/YGBgfwAOABsAACgAAAgJKwAdADkAAAAA mwAlAEkAABEROwAvAF0AABcXRQA6AEkRERwcUwAW/ysA/yEhbFkUFABRAEcaahkyZwBhAAAx/2EA /1MgexZDZy4u4iZZFlFGBGguSQdSj2oYuJAjFQBT/6MA/2pKEnUzbEpBmjdlC6QsFYMfsU4s/yBR tghkkm9WC1lDrTZyErAzFwChAHdfH4lHcbBDHLctfQCGlXpuIyafAHOpAQAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAADKAKxbASAdwpRScCSqTAqUiTZue0R1kP8AqABx/98A/1aRSjRI+MwyguRBcGjK ATa8QgCa/5Yit4V9MyW3jDZa7Vz/AP9IACKbokLPTcJYUiDTlaUk4HNWtampANBvPGefWInPC/+s AKcu/uJZf0zcZ/8Y/zp9/7HQGMf/AP/iAN+aPVaBn8ZDuq9xiziiydFTzv+aZUbK2/9N/8jpakze 4P+Y/9/AgunspfX2zf/Q/7GsWmORriJMZY1OP1BwcND////n/2lpaXd3d4aGhpaWlp2dnaSkpLKy ssvLy9fX193d3ePj4+rq6vHx8fj4+LLBZoC/eMbw8LKk//+z/9GOo8PcN6CeVHaucHiewYNkv6SD 09E/Mv99AER4IyRfYA4OLL4AAP8fADE5ANmFPgJ3hbDYgVYhHQAAMIjIs6B5AMDAwOpwgVHxaf// gJF0zf98/6L////78KCgpICAgP8AAAD/AP//AAAA//8A/wD//////ywAAAAANQIJAQAI/gD/CRxI sKDBgwgTKlzIsKHDhxAjSpxIsaLFixgzatzIsSNEAAAEgqQIcqTHkyhTqlzJsqXLlzBjypxJ0qRJ hSVz6tRJs6fPn0CDCh1KtGjMkgNvHsz5UanRp1ARhoxKtarVq0CdTl3qNCJSrGBTbsU5NqzZs2jT 4jRYNmlbi13VymVYlue/mzvfzt3Lt2/Ltm/jZtTrF6tdglu17nQrUnDhx5Aj02XLlrBGx5KHfnWb OGRdvHY3Zx5NejRgyi4tl/6bt3XSxnU5V15Nu7bfwAVVe9RtmyNmxKC1NubKtLfx40a/4kY8kzfy is6BM899l2tl0c+zaxcL+/T05ts7/nYWnbe6VJG5W7sOz749dMbmX8unGd19wqk2OxM/Px71dfsA BngfevPFZyB9Ak7k2Xy/HWgefvdhl+CEE0L4XXz1qZRhgvkNx9lhsUW3IYUkGufZZwsSGNSIANpU oGBIjbeYVOuVaGN2J6KWWFEsumehgRkSJuRiDd5o5GP4CUmfckr1aF+S8HWnnmpjFfehk0dmCVaO ERY5WJP6qaglgVVKt1ZDXF6X5phs9gUljWKiBCN1Dh75o4q61VflYXdh2eafmnnoX52XYQZYimMi iudkdML2moQ8eQnopDwWeGGfktJVpF6ZPhnnneeZeaamhFJqanILKYmRpI51yh6V/oxaOqqmfp5q K3cD5gqdiKCKeqNlvMX2UJIghnnrsStGSNZ7DsVo5aAknihcrGhC+11xtSY30p5gooids8QKymR/ 3l25bXB7ymYjp3kyS1aab7bZ5KVLoRkuvfARm22g4l6omKD9vugvg+WSabCYZcrKoaWbkrTswHHa mXDE9S6rb6PWpesmneUimqLHGPq7o3zCPloxyf7h9SvAferKYmBxeTvvuhi3+/C9+FK3b6UnMxgy hDuOjPCD08G6HMqNfrxyy746qnCzxlpLtNAUavx0zsBxW56aEm7c86JBoxf20CgnDKyswnYsdmnZ wssy0xR79RunD7oaXsGhzqbg/qFrQ0Z30n1bqDTZBpv9bldpp9y3ZF3X9Oa/WA9bLZxXV6ig3nBl vDO/SF8beOGLCgwwqFQfmDjgb7vZ665rgomxw5jm3fPmq108+ddy80d7soQD+fnUofc+MdUlfyp6 77e9Dte2dSuPJchSey6xoqnq6u68u6+IOHkuYqqc095nTXB/6sYO/mbPop/VYG/vyxTOET8fcvRF S1y58r6ldzCyJTZuL/X3kxufqne56pSuVOnJXu0+oiG9KZB/kYkX1FIXN9aV6lsSGVnXnAPApd1O TtZBIATbUzwCNmhz/ouSCE8Gq2E90DRNw1/+MjbCaDkQTl7aXbAi97U5NSVR/sM5G65wWEPL1Yt7 FcQd+/63IZh1KINaEhwBU+UqvoGviC1SVviSqMSLvHCFKZziyg4ow2L5b4COWh0WtdOumXlFPCwR EeyM5EbZPYsx+poSeSiYJWm971zcWxPbSCVDMS6xgZKDIh1LSJkzqlFKoDEgH5cmxfkRLYi1aWEh qbWqOL7xk79iZCHdNqs6DY5NeBPanR6JpFJykZOZW0kTQUmzTZpyj4bcnt08BS1VqvCVe9EkDz8Y S0Q2RY5/KuFn6sdJAHbQTj38EOZWGExXUpOYcxyivX4pOyByk0o75FgjkxlNLpKxMMIMYCK9KMsJ AmlIk7Ja5MLooCHt8lXl/qyn9BhnSGAycFUPBCdzuBVDVJqsi08M4dVO2cd8WtGSfoPlNV14Ges5 rkYeOhTz4gk8JebojqZUFiuNKL40CgeXEe3nRGkFRxA2C23D7B/ouhm1ifFQfVE8yRcRJNFa3VOi hxwU9o5nUJOpCmo+HJDKcjrDlKrUTz8lZlQphsSDQrSoeLRmAavXPaYGFZ3htCUVp0orso6uqlm1 qTd1plJanumca2Rbq3CXw4K2VDwzat+C4EpHiyo0d4yqZFyREzO6wq2bUtLpENWa0ZhWTavTJCQT 6TfY2jE2pBQMFj3XaUx2+Q5QQqxcCjerM1FWNpNdZean8NbWrcoSkFkz/i00LapZjI51a449LVi5 KTZGRpKzxZRJuFjb0FLubEqH86du1ZLXp6WWXSD1q2ubEyPKetB692zuRZeLpDAe9bLECS07hXI0 rE6xjgiL7kbUy92wcNCOuU2iT6uy00G+9LAhOuxL2EsV/tLGv019quL+iVjA9he0b1RZ0Mzqxfo2 WLvHse1usInZlSLQSQzWpnknCLQr8g4qetxOiBX7Lv3lTU8DvpyDsykvFZNpxW7Vnh5h/JQZZ1i+ BuQvX1+HU0VaGCY0zgyKpXPjCcvYxkU+C5IBjFQa9ThUyGTajv+jTp5yFLj6jUqQY7vkLnv5y2AO s5gh/MPuJHaTI1Xt/v7KKlYrX/m+vAWxZpYsYjp30smQnSSV1SzSLv5ky0IGJaDdNWfksnHE6w1u gaV7SS0elby2muWPEwhmRRNFPXXGbYB9TEXJjnNU4v3zoPkpQAv603acrjEaTbTqRO8txnG2a/AW ncHsIkvSr26z6UY06jLjqL5Byk+UHTvXV7ZNz5E9VZAITWwTp7q9PBtrvnzNqr9dU0aZAm+ylX1M u1k7l9mWMLR9ci5K51i2ePY1xNicwCZn2cwj5KC33GldhlHzjOP+8JlbjWsCExWg38KWuZK8MKCq M6m5NKfB852ar8LZ35+Vk3rJV9mwapE/tOYxjhfOcE/eGeIE1jaQ/n8bV4szenG6zm+eq9xxl9Zk uglGM8F5O/MA5ZFI6WQr12w80E6Ht+YtX3m6n51qeCKbxfU2FYoofcc8EbTUDSspy4MOxxP+ccjq nnVBaWdy/l0WupuSFsCZWGyqH8XD5rOSE3WNSR1tW7kPF7rSWyvrWNeawhm3SnXH5WFwBbxb/+l1 3L2naYU9jqKGJ258fT74W9M9pj4d9sXREiabQi7ilk+Z4PEOct9pVmqIezyWOc9txp/c2XfvfOGV LD2VS5JsaquZYT5+qRNC+X3WnS/MjwVO22J9sntb9eZ39emeL85wqEO93jFc2D3LGnqKh/vpOU6p R3Ot1rokregn/s2jMp1mbISHaQWBXtEqknyhCb14B0O9+87fyvrTv22Ro15Nw//us7GPv6ilaXpf OT3uoSd3k4d0vKd/3McajTd7/zZTxBM/4id9S4IufuRdU2Y+lPNN2/dX4wVB7Odn5PZMbKdqC9hR DSg8A1WBkGZHIHgliAd6Uvd2/Ud7Xqd/w3dWAqhlFpiDaWdSTRdI36N9+oYv9CRBZUVGGpR3F7hp jjeAkMQ5seWBZqdN1cWEtcdrG6R1EFhzNRhob7VH5BeD+8NkUWhkLwaGVFV1aBNsHldDvcdCX6h8 pQWEY2hk53dyMLNealVtwLZGScVaXziFGDeHf2Y8CWh0ivZQ/l1ydiUnT6bjZFyHXnAoiMKFhAuV XK71fQ8DHoM1cVhzc6aGZpIYhO8mXYL0c+91hvCniC0HV2DXiqOVhKHYE+iGWCLnMKvzTL22hXfT aEknYTl3gO4FSPeidmb0d8AIYmlmWHUHRY2zfqIIbR1GiQpnXLoohWpWgphnd5R3dMoYidhHKL0i Pw7XXoyYW794LdVoTJ7ndj5zjej0JaNka85FiMQ3jty1TFB4UyYUYQcVjoRTMm8IZAClTyUmaLlS i6QHa8vlWWx3juVjIr/jHXyiTOmYa9dDZAnJjlC2eDDYfqcFLEuVkCDJRh1FkKSDPxUZfC8TPDem Y+1oaRsI/o3bJHmgt1HPoTQH9CN5iDy3wWtXVIrMUnbJaIbMtpDdJpJPCIFO1YgvmXkBQ2oxZ2Jk BV5HmDkolG/Rx2XldUQhmZINl4bBgUcTqYMRtHuAx2xHs4L/Q4bk6I3yt1oQFpKxeBU+WUZ1qXFt 9x5cN26IaIVkplBeOZchaG8GmDNLBZQMpEN8iXoE14BDKZgpCHEM1kIf0yM2eVf3aHyZxZScOWsM BZl6V3THxlUoGHgaplveF2vUUzr4eJjHCJqqCIBb5JEYuTxqCZMf2Tmp1TIAKU8JFY2wqYCyuYz5 CG8v95o3GG9W9ZmXeX4KFpHIGZz75W9bSX140kSlSYC5/nlJQJOa3vlO/dhot8k4zJNXYVke5klx U5ie4xmZH/RtGQieLfhFgflf81OZJ/gz4tk9OGlVtlF5RFVJOrluxlKVoyhncNaBSRc/YceNMZmZ Y7NXJAMyp/Q9r6coAXlpfBYlynSBJ+mAU7c+ieSQcLd2NzOdMlmVElqGUqST/Ikh5fhfpxOetWc8 gqWRB0pfI2pwv9duLoOi5JiaQaRBg0OhB+Mxs0ga6CJOFDktn5mUxIkq99V1JSptKRZH9SlX/scl +tGi/YikNUqSvFiShJg4j6RGWWqOdRWCtXWlr7WYorJXLiqexodWJLSh14mXmkePBZOmLNeedbNs azmC/i7Xca7Dmxgpl3BzqD5ioBDVoS9pSZmnqMKZkQ85ekpFqLvhp6hlJs0ppwxTU5z6F2lnRi8i ger5gjMidtuIqcX5qiWVnbZJdWbam6wZNdEpnU5onUr5YxVKqhmaKN2ZNIYSprrqNe4Wnxx5Zrma qaMKkSgHNrPyrMeagEqFbwo5k4+pl806g9Gqn9ZarXMhqJdqYcxnoRL3rWNIk4Upru6lkJCorIGY PRZKrYQVrgvqru+arIDpqjGnQ44qiafor/rKepaorZYqOfMXL/bKj/iar4QlbD5KVTbJnt1aqfyn kob4JYj5cvG6rgQLsff6S9OCjq2njVDpoJMBKWIx/pkfKrA9uk0Fd3zrSJgg2qtyEaMAh66nGZQi i5rkSnT4tJw1u3E6u0A5Ol5/qYRzlKTQeJcW6SOaCT2PuqRuaRpaaG7t9HFZ+bRFmXp3WjbX+WT4 J43vuLUdaY+f1LX3iGF42LABFo0fWjx9+n5U2H0VZqhQtakiVj83+qgDQ6kIdrf8wrbb+aBMi1rD 2ITn44MuIocylYlyVp0yqbaWW7BIK7l4a7iLWFFvirnQaqWowrnVx3OuhqWmC7p80aY1RpBFRGM1 CLeqq1xiuF8BiEWwq6GzO66lVmiDObgCuau7axZWqKFOW3rB657DS7xj1rzOO2MVJ1zDJ7vLW71Y /ipx6WNPtZt11tu9LWJnojtw4jaE4Ou95isvkKKFunS+7Nu+7vu+8Bu/8ju/9Fu/9nu/+Ju/+ru/ /Nu//vu/ABzAAjzABFzABnzACJzACrzADNzADvzAEBzBEjzBFFzBFnzBGJzBGrzBHNzBHvzBW/K8 IjzCI7wuJHzCKPxlILysmftYNrfCIZpJtZRFMHyxyOrCNAzD1EtiNvTCNbzDbElSvKTDTCeMzceh Egs/seN3TGzE2AafMytA6DORG7WqSSxs4wK5LAzBp7NK1oJ8i4qP3AmdgGt/kXuJTcmhkYqN2Ci0 HAyfOynGXkyzrJlPsvezI4vGN4uF2WjGOAvE/scKxzWqcrj6eoYsSjPKkzZcltlkoKrilGWrqbxq wXDcdKKzmkJ6yCiptSwIq2KqsVCqvYPMyeH3tR0syPWUPup6hAMqhOKEpzjbG5ZZppUYcdE6x+JI xA5lrADajr/qyoY5j2fsY4wFkNGEyDG8xQ+MyrZcxl15f8hMoMLcw1vllG3sj3csq3Z5lk08geqS xVosnSZqgbDFg2UTloFKzorxxOVzuzjcu7xJxVojsVWILXU4epA8xnT6j2vWR79pxN/sxFfpw9Q8 xA+4k/tUzNUnzZ55pJgptQ/Kno7LzmlU0UEctvTSxpyp0Yscuouaxjb70BhdlMOzbpKap4V8/rk3 +cX/Nm8LCMjkVqcgTY8iPbTcGqbGHKlDg5C0iSMsjTzRvMctJrbG2cz0KiDyA6k5HTqMmLUEfdDG eoYPSE5EnbfH68afXM04Taw+GFmNidSzY9Kd+dILTTB+PNana9DYh14vO7e0yMNqnbFyrc5OpFFv 1sd3nMwZuXRLLNHtpjWdXKoT6InJqaeJfNJTjaFwTVNXXIx/HcrEOFyh0Wp6LYh892J1PbGJm9HM dKOqxMZrPNMXhpbHA8mTumCVLXR1q88l/ZSIjdccrcwObDR5zZO4rJGgLVSF3c5ObC5S19vn1ltH vRZ0W9trhs2+pNOKnNqzzVXGrdAsI5Gh/n3c8zTDgdjSnB3duF3aJtiuFBxa+dzMuyZ+2RvZ8yyv K/2j2D3VH80xEy3XlB3Lzd0ldVTcm2ySYU3dYGTd17feQt3PY8rPkbzbEmx9Gg3dYNw5DJ18ZpvV hh3V0q3gAI7X+9Tg301b/v2Pyc3HrR2G6J3HvZThmLfhVyXgHu5KfznPJbu/0MXb60zKrJqxxtjX 9pzS8i3D+9G4Lw6lwQ3ZfN3EjUuKwOzQsh1F/KaVV62dfTvMEK2n34obME2XQ67PzD3JsryzU8zN wB3Gw4W2cT3WznRE0TvKmxxQYK20+t3Q3W3jad3kEB7SfMyBGkdcgNrT6R1UsWdarlnl/h/ewiE9 QBoVrEvO3dnd0U+cwohOwkVO1k9uo3lKquKb6JI+Zmd9VQjN59f1ThjEdIuV1e5T6OnrpOPt5VFu kLr9507Ohgxu1L+boJ6e1nn+ylW96NP6yYF+loLt4nwo1pd+4yf26p57s7ENo8sN67auus2V5Tpe rodU6tTZ5vmc22Xo67Xu09SuwNrcSpj+d09naOQs3GvY0cLp7KCZ7Wd77UpK68yF6djO7mnB07vo 7q0q7gJs7skj71yI7vOektvLl0CYk/TMd8WIxYDnzV9cxWgE72CVbRkd8I2Nqoxb0cJnmoSXf6q2 tHqL8Rxu0htO4nNa4r2p3Arfkxov/t623OEdr1pdlUpqLOz95e20CvOrDtQmXl637eh8Wu3/KfNi TfOwnd9g7HpLzUIvD71Bh2RuSvNLPcePrtNUi6OSTJ5G3/OsvvTdrV91zIpSTe8PVr5B2mUcDui7 1i3n7WyWPOFFHfUR5GWKrF2YTPZchjlnH+BzPbaTfvfxvfB4v/eCC93AnOCyzukXtPUVz2+YZoqG f/g7B1JiNuV+WDGAD8uhcZBEr2Ve/7RIr9ziTeeVP+Bkrvlq7zfl+3WFXuksiaeuF+eqf2TihpWI NvOb74K62fJb2co+r+90+fpU7/dnbft7jpfXDPuFVvKYr/EADz+BVMRBzqoznuu+/m3clrV60H/Z P7nO1F+e9KzrORiX+O4beZ/xpGXvN0zu9L2muE+83Y+H5Y5r5P+erMZp7R8r8R+L4r+66p7u6b8l +T/A9V9/5w8Q/wQOJFjQ4EGECRUuZPgPAMGHDSVOpFjRosOBES9u5NjR40eQIUWONKhRI0mUKQWe PKnSIQCYMV9GjAkzo82WLhOy1NkzpEmfQYUOJVqUIVCjSXdCDEpz5dOMUU3mLMpT6dWCSLFu5drV K0aoX41aFdqS7EOyStOKHaqV7Vu4cTu6XWlzZt2bU+3WRStzpt+8fmvu5Qv2L+GwhptmZQoV8dW1 dyXvBVwTIk7Bg7NqPmwWZ+DH/nTljiZdurHiqIkVu01rlarWszdP54ys0+xlxI/Hnp7NOPVq3q4P wu6N9zft36aVL38rGixtpMgTCy85nPHt5NWz98Su2jdk3qmhi9cOnHz58Gi/o/5Olfn7lJ4t8wXc GbTe55/tF+7+mj1Gp9gjLsD/CERNvaUKDI+79br7j6jaELzOsfKAItA/86RSyMDj4PPQNvRYC65D w/R7icSp2htON6c4NA9B5+ZTTULJ/rJuOxBHvAxHCBessTEYQ3zqwhUJu40lyuoj8cb7mnRyv7to 0m8w/Nz7cMMDfeuvuBFja1BF7aQL8ssWfTTwyC8V5DG+JKXkScmkIvSuTAGZ/qLTu+fSy/I8MJlM TsQ/UZQtxbAutPJKhGjcc0tBZdtRy0crZLI/F98c0kw/AUwzUR8RdUnOGR2V7lJNccRQQx05xbPU 7DgEtE5UCxXS04ZepXA2GW3cbNfAShr11AxLXHAt6m6VNdBWO6UVJVCn4xLWUPnMc8lgGa3OwWmp jS5S44DkdtmFOIuWUezc+3VVa/Oir0YkWQyOMyrpy62+eDcF99Nh7bKsMtyAbTez+ert9UlqtYUU V3rhnLdfX+/F8tlsjw3WWS9TldhhimrD+CdlN1a1wFzLBVPkTIV98FBw07V15ROHJE4qupzzOMFV Z56rY5sh3nPRPiOztOCJ/jeuDDMZBWa3r4AnOzpphdfM2eKnPdI4aqBVRvbiNy182WWnqf5QZq+D Druiqb2OuU0ojR56P6Rd1kuwH8emFWyzcZb74ZrrzvtuvsWim+qy+/74wbD/Fvxw8PaOOnDEob7b 8MYjL6vrROd1c+hcZxQX4IsYl9xzX+WT9+3QNitaX92Oolxy1nMkPFw/K975atBnbX1wlGm22FZo TVb849xvF34kyEuu9mpWnZWWbLs/b9741pBXNPrXYf99eOylXh36VKN38OeNakdcfExp79JeRZm/ Pnv2LSoevT9DDjl+JdPP+PnGyTe1339lF5fU8G2vfQOUyPvWI73l/c9F/tVTnQDHh78DKu9Y8jES noKHtwsSUIMYDKDtfBezqmnPgYfTnwW7V7CtjVBSG2Rh51SoPOp9sGfwcx8ESWhDZMUwhTKMoPoY 2EIg6qxW8ELd6UinrrUVhiMl5FsJ6wUwzIEGiUlS1xJfGEQCGtBjTHwcDjGmRSxqEIxf9GITy3iv MYaxfWlE4xnlxsWUXVGN2WNjHOX4Rjcuq45zFN4e55bHwgHSU37kI+sIiSg4ji2Rf1xfIdd4HypF UpKTpGQlLXlJTC5Sb6PLZCc9+UlQCiyDjmTfKBUpRuyZkpTDC2UrXfnKUIoRlrOkZS3htEpc5lKX u+RlL335S2AGU5jDCSRmMY15zF0GBAA7 ------=_NextPart_000_0000_01C71122.93BDACB0 content-type: image/gif content-transfer-encoding: base64 content-location: file:///C:/Documents%20and%20Settings/Klaus%20Holtz/My%20Documents/My%20Web%20Sites/mysite3/THEORY/info3.gif R0lGODdhPgLBAPcAAAAAAIAAAACAAICAAAAAgIAAgACAgMDAwMDcwKbK8ICA/4AgYP//wKDg4GAA gP+AgACAwMDA/wDP/2n//+D/4N2cs7OP7ipv+T+4zUiENpWMQY5eQqBiemJPrB0vvihmdgBFAEU+ AWooE4U5akoyhQQEBAgICAwMDBERERYWFhwcHCIiIikpKTAwMF9fX1VVVU1NTUJCQjk5OQAHAA0A ALeZgYSZtL29kH9/YGBgfwAOABsAACgAAAgJKwAdADkAAAAAmwAlAEkAABEROwAvAF0AABcXRQA6 AEkRERwcUwAW/ysA/yEhbFkUFABRAEcaahkyZwBhAAAx/2EA/1MgexZDZy4u4iZZFlFGBGguSQdS j2oYuJAjFQBT/6MA/2pKEnUzbEpBmjdlC6QsFYMfsU4s/yBRtghkkm9WC1lDrTZyErAzFwChAHdf H4lHcbBDHLctfQCGlXpuIyafAHOpAQDKAKxbASAdwpRScCSqTAqUiTZue0R1kP8AqABx/98A/1aR SjRI+MwyguRBcGjKATa8QgCa/5Yit4V9MyW3jDZa7Vz/AP9IACKbokLPTcJYUiDTlaUk4HNWtamp ANBvPGefWInPC/+sAKcu/uJZf0zcZ/8Y/zp9/7HQGMf/AP/iAN+aPVaBn8ZDuq9xiziiydFTzv+a ZUbK2/9N/8jpakze4P+Y/9/AgunspfX2zf/Q/7GsWmORriJMZY1OP1BwcND////n/2lpaXd3d4aG hpaWlp2dnaSkpLKyssvLy9fX193d3ePj4+rq6vHx8fj4+LLBZoC/eMbw8LKk//+z/9GOo8PcN6Ce VHaucHiewYNkv6SD09E/Mv99AER4IyRfYA4OLL4AAP8fADE5ANmFPgJ3hbDYgVYhHQAAMIjIs6B5 AOpwgVHxaf//gJF0zf98/6L///8fADE5ANmFPgJ3hbDYgVYhHQAAMIjIs6B5AMDAwOpwgVHxaf// gJF0zf98/6L////78KCgpICAgP8AAAD/AP//AAAA//8A/wD//////ywAAAAAPgLBAAAI/gD/CRxI sKDBgwgTKlzIsKHDhxAjSpxIsaLFixgzatzIsaPHjyBDihxJsqTJkyhTqlzJsqXLlzBjypxJs6bN mzhz6tzJs6fPn0CDCh1KtKjRo0iTKl3KtKnTp1CjSp1KtarVq1izat3KtavXr2DDih1LtqzZs2jT ql3Ltq3bt3Djyp1L9y2AuwAU4h15dyLev38x9q3rcG/evYSF5kWJmOtghH0fc1ws+SFlyv8q+12c WO9lgZo76wxtkHNG0lZRZzbtkXVFyapFa3z8WXbPxgldW4w9NbJv0IYHB199uPjAxoCFKzcO2fRv 4sdp+0aMnLputrhXa4cOffp27sNt/u/+Pjwwd+KXi+P+HV7qcunal2cGnhw4+c/PAed2Hpg98+7x GSecfXRVV1567+GH4HXiFUZggAT6BxqEz6GHX1XK3Qfhd/ZV+GCG+f1XGn8BHhgdQdN5Z5t1HIKo YIcvNkiRfjBGKCB8EtYYH4YxUljbiRWat12OHhYE240o/nckh/QlxpmL9MUY4pAiymjZiTpC2aSE OD7JYFNa+vhgjWFCGeaISW6I5Jj5zUflXBmKqeWUZ1rpIJZTUkjmmhsyCVVy6qEHnoElRrmgisHp Zl6i5xmI4KDwyRVodyn2p5916Z1nJ0OYLghplNE9uiVtm5Zq6qki8TYhqqy26qpE/rwd9uqstNYK qme25qrrrrz26uuvwAYr7LDEFmvssQu5hllpqY7H0pdpCXnaQdDm+qWyKxpZ5Wz7QTYtrJxWa5Zz F2GLLIOVqdqWrCiu2tG13grmV0PikmXujMweyy61+Xa276qKTkhqnCqm6WWTYx4XapKY/UjlwUCi ubCbXWEbJ3kPC0wqm+rK+G+HEc/nMKZYaoqhwigrO+nK7i7b7pAAp6xtzC9TR7PLyzrsZs5imftk zDbvjDPPO9P6c8o5s1y0yDcvfRWNNBvZ8tQUu4xy1QNOXHPALROKddYUT9zxyVJXLbbMUQt98axQ m+22lxcfWSTZV1uNttM4u1t3/tNp653u3jVTDfjdfjqW79FE4+233iW3ytrRirs9c4QSp1a204ED TLRhYWfKM7s+My4r4k2Drjm5S9dL99uRRd1w0Zv3jerjSIPeOt+mw864e9Kpl119nC+p846g9s7f pAZHWjzEIGtbsOq8p7svoJRWP6jWjvce8o7Vib2kyXAiaxP04pc/mfk0ZYf++u+Ozf778Mcv//z0 12///fjnr//+/PfvWH1hu9JrmvWyX8HrcvKqlfoSqJbPka9wAVSJ+4xGr4RF5IGbwqB4aAcufr1E g46LYL9AKEIF6op235sZxLo3wu9pj2MBbBjkXPWx0YGnRcSzEeQWmEFqkUhh/qNDnYjaZrg0pa5s WctbDBNHtdgJLXl2o2Hd2kS56jkRZiW0U+hShygYJY6ES+Gg2sQIN79pZnisY5gS7wZGwrSNdILb XNBkN7sRvVBycIzcV8i4uya6bnc/HBwWBanHV/Hxj37EXOf6eKp7IRKPcZzgny7HRLUdMXeZ4yDn jmhJNpqwgKSbHtpid8U6ApFvkERkG5lCKBahCW5KWk/BDIYwGB6PUau0i/ZwGShevkmHFiwVAFvE y15eqlL+S6Yy31XAZTrzmRBBHTSnSc39SLKa2MymNrfJzW5685vgDKc4Y0K9LK4kUeji1gepMswL wo+IH4liLoHiQJNUqza5/kTXde4JF2laZp4UZGS5Jte4rIiRL5y6mjrj1a/c9LOZAnzfDDfiSIGy U2opbNftakkplRlqk40qGXPUt1H5JKyYIgvkFKNoFGkOjERl1Jg5bbXDn2XqQxmDXU2x8rhSWjGR 8lSlHIGWxMNd0W47vCHmiuoUqMlwqGYrT7CE9FTcKRWfTgToTg4pLatqDqKCSqXP5hjUrlntd548 nSa1mj7RgU2JcGTpCe24xlvlEa63Wt3bHMrJzpV1cGXl6ghRiVSNAm6fRgTTYRMZuUryqqITFawq LTpJUMruqKXL5GL32thFAkh3hX1dFQurSKUktbOf5SRb/XU4oH4VtK6d/mkYd7m2UeW0UR7t6PUQ Jp/2bM9SoVIpuTKqW8rSM1LUY1FMTeQrkt6UmEA8GMl2O053KrS6dVktdpOFwO2Gz7slkaV2wdub a5L3vOhNr3rXy972uve98A1KOcfbmg7a977G5RFVW/NGvbhzmAcsIkOvCxKAAkqu3R2oTptKzmBu xZ/1JTB3E7o4sOa3sj4kqIIzzC3HTvh8Fj4KfSM64p9AmJkhHvCAIZxOrXwMrPn077ckfCeKYlSl WkNODlOKNcrBJsc3jSmPWwjTNtlsYwbN3JBxaqmD4m1RuGVx4LwmWwZ7VLoC49hzo4w0Jgs5gpok 5kQ3PFrR/VGqP51a/tCYClWbgq3Mmh3km128qCATta5mfmsVl+o8gqX1aQrFpxetStq9hnJAZ21o FwvJwCai1bJ9/eKf5alnzoI5mKJ9ZIlHo+Sa4rWhkWYjZOPMpE2nT1p3NeuPQnNaueXtjKCGq5PJ bOnuxjW2ls4jYRnryL9tVI+mxokQN5torol62ImNIalxCmhIo1aEyotjls3cNRgjMdBjdhaxr83Z TyOupKPM7KcvXVrS2vDCGAY2KpmmaMLJDJPKDjcmg91gSG8SpKTMc7d3nWhbwznbM6ItGuVmW+8E KYjGHNmBgjxLRrV2YQb/dWpohFIL/ZiuAu/tlv1a55FeXL9A7jHy/l7oXO95fFRd7WkZY3ni+JoW XzSWMa3rVy9+upzBMEd3imucP3FBi943D695wzqvnO/v56gBetCXzvSmO/3pUI+61KduTwCqDow/ B7U5rx4SDM6a56fuaoHHbnSDwtMrI64nvUjoviSukjRKh64Ah47QmL99tdeku4hTrF3oxb3uMde6 daMpuJnrHMSBH6ywE3x4vka46AJmZN8Jr1iN4njJGBNeoUqKcJkS+ZV4Up7pfi3DZEvc4yNHdCD/ TsbnSm9lL/WyX/2bUdvxODw/RqNPWl9LVt8R9SK9LbNIv3ocYv7slMdT39h81X9j8a1IraSgO0uy 2y0aqsqXXLX7/vosbo+b10IVEB2df9Wsnm78PAnzs/nNVJ16uOCEdnvzGx/r2h1SqagVZa4JPHD7 q7rC4yYqkRVulOR/LtF6ZwVrrSZtXJR0KwVYQMMwK/V3jfY1ZuRkqfY13/dwooJ/GQhwycdYBRhq hqZ/GUhuJ+VuA0iA66d9kHQvunaA3ieC07ZvsWVzJFiC+0d/bRVnMShZjBaDMxhva1SEehdv3GeD cdV+Dcg0YyZ9WxKBTCRpSshRSWUxNtgSLiWFfbSFTahaAnVrjjaFDEiBZHZr8EaCoaSGJWR+ObiB uyFwmOZqDgduFkg8EXc84JNDosdwSAI8ayNxJqMoCcdsKdFO/oJiPLFmUtZHJDpThw7Hh271PGYY hynXHpcogbqFXBFXao6oRsUUipX4WIa0TRTYcs40ikmmTWaIisnEQ1SnQLAYi7RYi7Z4i7iYi7q4 i9w0Xy11hAQUgsIYdrOIFKoobIK4UMqoYuNDVEWBYPLVGzt3G5D3YAUFdpVzGki3e4zHjawkjYmX ftW4iuHoQTK3jOa4VTeWbJ7nZ0Gkib1HiB04h0ImftH1RDDzcSxEVrH3eVVGEi41XP9SfbXljnro ZaLXjnl4Ud6icPFoVMfnIiPVOJznSjDRU+u2ZtGlNGJYaIE1hCllj2jmZknjVdMXkm6Ffue0X5Zk fWnlkfPn/kB4dX2ElG6a6IaKZDtUuJM34231dliP9mwfeEsjKI/+5lVgqGmAVGkn2GU1kYAbQ4Xs RnTgt0SVxltVeYwA6XhUBX3SY2/ytjVMaYDrBJbmGIB/NkhF+XDX1ZEEeILmlpX8x45l6YJJGVU0 9oFHWYVp82I2eVosiFgL6IzIBlhI9pO49m+cN1lJKIRxyZdJiW9vyYU7uGwXuVl36U/mZoc4eW48 mWkM6VmRyW7qB5fUJkcvBo1Vt3K1FZEoGZFf5nsbR5ViVnDG9CaFaJGJCHueQ3A69kGcCHHAB1G+ h5K6yYiwBIhDZHYHKWax5DwW8imEc5w6GW1EQT5ayYuO/kc/rjgU2KmdnHY/3SlfHVOM4LmS2Skb Ynee7Nme7vme8Bmf8jmfToJ84flOsKZM01UxJ+GKD5SewwKgpViO6ohf9tSN/wleAipFBHqfwxhe lOR6Rkllthdcw7Nc5cM1QyRc/7grQmRwckearVN71vmQ0ZmPwLRgFlqBmEWE0qaRWfh863NlzPeF 5+JUG1k6blejPuWGJIk7HNmg53iHzckmmOmBefVaM+pB3/Z75uOFrcaPhhl9X3mTIkhWVUlrLUqI ZXhseUmX+sKkKlh4YRpoYmVrJrh8uJKZuGaEMbZtafg5O/iRlnkupNaZT9pM04OnQ+ajp6mjUShu NFgu/kQEXEnKQtK5m4iKckJqLRR3TAo3S8Wyn4vkQozonDAEncGFm1F2I012ojjHg/TpFuM5pNc5 jafaqKNqF91YQcaIqtdpnqsqKQQHK8AInLc6q7q6q7zaq776q8AKdSDlQ+O1oHtZIOuphayCiB1q qmHBPBD5FPd0eZKiqtzZliBkrDLhn9ZKjRTWrfYCrvETYAY6LipzKJ76SyJ6lmdUZ/cYfFgFraeX McM6Lu9mhf2BjzIlkKK5obLaQGbUp/VoZ70HQTyCmW75mtGnpjpKSjWqWhjJaGjRceAHoyPJHimK Z3Dym2MZma82Z6tomlPpgSSFaQbYpFD5fy4ohtih/njrx7FtirCsdaaX1bESC46MGVVpapfYZkdt Oaa75mwvSqouW48vGbORpK0mpng/mLMaW15HO239BodQCJJbmmWo6Wx+eRYw6G5zirS1NrNS24i5 I6doyVOaMl1R2WeLyqhsG3qwF48HR0uaaqYTW6udmnqduLe5iUxu5FR72mObiigap7T8Ka4QqkWw +njNGqzKiriAZyWlGk+L67i4aqz/ml2Zy1+5arme+7mgG7qiO7qkW7rYYXU2Zl/1io4fJhrMOkDo U1ERBbk4V3oSxJbqNDadCxYg6KrvpGFFt7th1KrBSLtnOXjqyboZioQ5Z7iP+670eDayd0NQxlH7 /mpER1akxPch0TZH1jh7J5eidwibNIWEj2pTCotDxOe8LNqjgOp+Pbl9VIqHrsOTpTaZSfg/UemT gvuj+busdptquJd9Iymq9FSUvyNlYaWXXvuWWOq0CKxGDva9g0mkneZojroecqmUQNpsRyox6iez ntViVfu11DZ70Ma+kbvBU5uWwtuyWotkZxujL/yUa+mio2lofFZuJjmV3nZx94aUaGdYz+eSpMnC AfXBX6t/ORw9wnmhlhqvVnS+FsRcXcKbdEu3spRjFQO4Ije4fsiom7sicgi4vZmuviSpprt4pli5 axya2TS5b8xONfykYzzHeJzHerzHfNzHfiw+/r6YuMQ7u11HyM7qxCrsMbeLL9/pu9Dodc4IkKpp q9gYT4nMGNeYusTicxSlQduijcAreGBnapNcrulYvGPxdcrbK1vrxt+afJM3yI37s8S1rsLpY2tL hwQlr8/TSUv2Zc1DrxVJke16SsbXfc3Br25mI7hlzCb0Q5S4XP3omjXIr2M7ZYgVv/eLXw8bk4wp Xe1Xmj2MlRpju5vJfuMcZjB6mcf7U+s8k3UMw/06tD5cfg2rp1XFRfi3NdVcYcjLOn9jV1AKgGm5 wHaFQBX8fijrzFSqe/wYz4bcpUKpspfcVKg2aRgDwQatp9O5tl3olIc6jkJIpntGhniUn0Brzbcm nM083LTZmIPbym2QOcMGLLYSfdImnJMIHbWK85u689HczII3eH6iBYf4m280e5dE7c85/Jgjjcyp BLE9aUPvB8D2NqJ4Wc8yKsTznNPGhpOz3EoF2YcwZT3KCapwm6m3RZDuCnopyFurJ5td6cqWrHK3 PKJlrSF7ODu/V6GJqmOFK75WKbe+I6nolCfGy8Y1bT8VncSGh5jSStdHp6CaPBONbasQDciZPa63 mp2X/cegHdqiPdqkXdqmfdqondqqvdqs3dqu/dpWEhAAOw== ------=_NextPart_000_0000_01C71122.93BDACB0 content-type: image/gif content-transfer-encoding: base64 content-location: file:///C:/Documents%20and%20Settings/Klaus%20Holtz/My%20Documents/My%20Web%20Sites/mysite3/THEORY/info4.gif R0lGODdhBQIRAfcAAAAAAIAAAACAAICAAAAAgIAAgACAgMDAwMDcwKbK8IAAAACAAICAAAAAgIAA gACAgICAgICA/4AgYP//wKDg4GAAgP+AgACAwMDA/wDP/2n//+D/4N2cs7OP7ipv+T+4zUiENpWM QY5eQqBiemJPrB0vvihmdgBFAEU+AWooE4U5akoyhQQEBAgICAwMDBERERYWFhwcHCIiIikpKTAw MF9fX1VVVU1NTUJCQjk5OQAHAA0AALeZgYSZtL29kH9/YGBgfwAOABsAACgAAAgJKwAdADkAAAAA mwAlAEkAABEROwAvAF0AABcXRQA6AEkRERwcUwAW/ysA/yEhbFkUFABRAEcaahkyZwBhAAAx/2EA /1MgexZDZy4u4iZZFlFGBGguSQdSj2oYuJAjFQBT/6MA/2pKEnUzbEpBmjdlC6QsFYMfsU4s/yBR tghkkm9WC1lDrTZyErAzFwChAHdfH4lHcbBDHLctfQCGlXpuIyafAHOpAQAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAADKAKxbASAdwpRScCSqTAqUiTZue0R1kP8AqABx/98A/1aRSjRI+MwyguRBcGjK ATa8QgCa/5Yit4V9MyW3jDZa7Vz/AP9IACKbokLPTcJYUiDTlaUk4HNWtampANBvPGefWInPC/+s AKcu/uJZf0zcZ/8Y/zp9/7HQGMf/AP/iAN+aPVaBn8ZDuq9xiziiydFTzv+aZUbK2/9N/8jpakze 4P+Y/9/AgunspfX2zf/Q/7GsWmORriJMZY1OP1BwcND////n/2lpaXd3d4aGhpaWlp2dnaSkpLKy ssvLy9fX193d3ePj4+rq6vHx8fj4+LLBZoC/eMbw8LKk//+z/9GOo8PcN6CeVHaucHiewYNkv6SD 09E/Mv99AER4IyRfYA4OLL4AAP8fADE5ANmFPgJ3hbDYgVYhHQAAMIjIs6B5AMDAwOpwgVHxaf// gJF0zf98/6L////78KCgpICAgP8AAAD/AP//AAAA//8A/wD//////ywAAAAABQIRAQAI/gD/CRxI sKDBgwgTKlzIsKHDhxAjSpxIsaLFixgzatzIsaPHjyBDihxJsqTJkyhTqlzJsqXLlzBjypxJs2ZM ADhz6tzJs6fPnzltCh1KtKjRo0gTAnC5NKnTp1CjSp3asClLq1SzapXpc6tXk0EXYh0YtClPgWf/ pUU49qtblV2H4mTb9m3GsmrX5p2LdidZv3vDYrRal2BbwoTJKka7mDFbu5BP8jU4mWthw5cjQ0Ss 1nHixlg/d3ZM2qLoxwVDNzZ8cGzhzJpjV6xMl+tm2bNZrx6tW3Xp3bCr7qbc+nBps7qH/8bNfGLw 5AqBSp8u+OHz5qnruibOmzdy4IOV/kPPjrmyzvGvsauPeH08d5DtUzvUGzj70vuA5W53b/b8X/Py XXRaa2It5ptv4sWnEV5rNUiba7StJ5xzSo2koHgEgpYcZ94dheBy3+33YXcUDfjeaiaS+Ft6KHF4 GoLGBSjhbRRmGNKF7qHmGXcHGvXhfr2haONyEqXIn2g9GpgjkSIleeKA98k444QlPhmhWPlhiOF1 ToImon5CSokkdN8xWeSVUjImWFpsZnkiWF+SCR6QUzJ0WXVvwsgeeFhGV5V/KkL4H44l4RdahICu aWhY9BXK1FWIPQhody/WSeV7mRmpYoE6PpbphCNuGqKlNRG64KOjxYnipKYy99yn/puaWWGgBcLK aYcBdknqTNT16mujFgZpo54b7qpUcLBFWRt7eMa646y3/viXkloaCxaqkn1GLKV8cmutssf2eet8 IEJr67Ec1qcYo25au1Kr4bW4V1+JTjZXWfbe226dZfpZYbJ7dmpfs9O669arxREKbLUGe6Xpks8K bKecE/9LcMNJoZkmrkVCDC/GRYU6JKYDMzupsxCTeDLIPl5KmsIeswwZnSX7K+tGrWZ68cyR0suu X/jy9XOkivYHcMXF1mjmxzLTpJ1pb36U83waH0xxqs9K6+3WDzNM683oVs100zdJTOPGHsFs8tg2 2Rvz1Z3FSCyXLnMc8NFkZ3Xu/tlRd6R2iTtLpayuKq8bMbUvo4x21F2ba3Pego/Ld98cXcj2ylEh dyiiPGL9dZaBL75x4zruLW959AZWNITgqqv6f6n/LFfdSNP16+2hiy5g1S3HLSx/WJs46u/R8S6q 8bYXTLmjYmZ9nJdhcryt4thKbj2kdrIdNlTpOmuc5sOVOb2A1ctL5/mdI454peqXuhmy2t/1Wvx/ Ig+T1s8/2Xzi0ZPfEv3kQh/cXkYzz9lNNQA8VcICNT+ixIUpuYMUkkDHufAVbVpDGwzuNgiUd/VP gOgRns/KU53WOa1081sY5GC3whuVD04lRA/w0ke8/lHvKtC6YQL5FcEWTi4l/jvM4QH3R7x+Hc6G QcxN8azjwzPZj2y5Y9FsApfEkbGvW94Zk4b0lzIc/ql2TXRiFSPTNe1MbUtAzBfR7FMzNUlKjWyc V+p4FTCzhdE5T8QY3nRnPSnekY61u9MfK9dDd+2xi49j0hgHSa4vKpKRUiskv8DoNYn5EZJeFM7F FvlHzBkLWXas2yUxCcTKkRKGBkPYyH44ylOS5EqSBJsrCZnHmd2slUxE4ywlM6wf7rJQsdzKvnIU TBIm8pcuPKYQkZnGWrrqhRY6i4P8o0LZHHKZzJSgIaHZJOdtUXpTMl0osylBTgqFg+jsSSbzJ6pd iXN55LxfMbViTnh282r9/jKiegRJyXiec56Z4+YrB7a5gk4yQ+/0pwOdaRlQUoahsVMm89qXT1mS MWEaq6dCcQZQuFQrRjfEVEKDlTQiadR9StsoVTx5zutVkoirvFZJvSUy3ODopCq9UUcH6lKLPpR2 r4RjX1RntJ1OxXI51YxR0+bQmL6vp0ntmC8b9sB/VpOWOB2n9+4i0aiG7TzxqWdVS9VAsi4LhpbT 3jV9ykWtelV5BKVOi8rKK/hZJntw6eFSb4lQPGopq00zHkThY9ey5VKgTgWmzliaPcE+tJiGgupb w2qbwxYvndOR6lcxO1agZgtRV+Xq/96KTcn6jLOd7ScioZZSPvpPsy+h/mZqcYbYqP7NrayFrWtf W8fd5ra3De1gJFGLzjSysE1A05fQlqvGC8qRl631rRJ1+9IqTbe60Z0qBKUjtdp2N3rDq6j6rkg6 v/12tecF7vHag7lk8TNBr2rUGZ3WK/OO1rjszJOc5Ja0k95JbEMq5MIWK0RcdqqArjUSe4/4XyqS tb6Wjek8q6nRwdWwcGo6ogE75F/HFXim4wKpWxW8YJjaU8NsTZCHcTvXXzlymeUlWWnv2bqV8Xd4 vWHVYE3T1ADPuK1/dehaV1nTuAEsPbU08Efpi7taQfW2XU1mjcslH/BR2Yj65OmT/2pZEWvVhLgV 8X9lldEsl7bHif0K/phFZyozY1e0eXkba6z8OfDOVbXtk+7rTowy+6lygCDWXYxjhWY+r5SvxZki 7yoMN5GNaFv4C1ahq2xaQP+ZQJeGsZx1tmKgDhm9kUM0i2k46gW57bSbA/II5xjnDNLYtG5WUKgy natK53erZCZ0dZVMPcBGk7EjjTKokfkwN9stzzn0cpqPzWqqVTB5/krh+54m7JDSc8spfa9tPQbs EOp1Xw0O5ESzPe4EY+eajL2sg72qbFaWW71VnG+KsxthSwHW1+Fc9vXinV5rX9eX/KbQjtXs3V9q W7sBp+6b1dtIQ//7sAOnJ3GLm9Ng/ziJUNYzvaOMcepG/KgFd+Wk/mFNUu2ieuLD/DLKf4JwfHv0 vgqlW70xuvKU6xtOeL1JYZeoZgAvEI8sdfkzrbNznee8rmdFeqKJ7rB+DhrQ876jGdFE9dvdeelW tbk8QwttwY184VvTuNR9bDvAlPizWicteUAu7NvyGpPUXm2Rb652+rJ9nG0OOjlHiWZZ1517Qje5 xZNd6iYeXHHttvXflZ6xrsr77Z08pjhvunjAB/7HuvY31DW/QnSf+eeFt62bdEz6gkbw40yddNpp fni4X6r1VO5z5Qmf3y5py6eof/k6ScnenY3004snXKW8rLDLN9yDvzZyvWSb7tiakuN4/rvwhUR8 Hj+l5k2+Z5yH/si/qAeV7ieGvcOTmh+Qml2/u2s8zOFzYZOKXcuY2bico09a8c4UzAgGeu92H8la F/H9JSd2Mjd+nOdPWnR/XwN24fIuQwZQ3QZMg9JGGHZC4Id7fTVzandFYWd/Cshz0OVSj4d59oVl t2Z8mXd8lLR6wxZPihKBg2IezbJXP/VuZjNYiddxyNZOK0iD1uZ3zzd7DtNxjhV6sQeATHc47DN3 64delPeDQChM4kc1olaB4UOEtzE0zcVcMmhffNZmXPiEUMg6XjiFBFiEO8gyU6d4o/aAYNhzL2hd iFdtRGaFegSH/aZqRtiGWaeCMvJpY7g0p9SEcGZJcqiHTjFb/lhSaDvFhi3Ghy3FcHdohmVoiPqh Tuy3hN+HdQ8GOpDIW6NDf5ToIdyVNiEnP0dndEtHWZfIOKyXe6HYTMKFVdgnVzxIh4TUWC6mfTME V/G3h1vITL7yhbqniyg4jKB4OjAlWJdXdZSYiwrEf6Qof7PYfJOYTBKoirFVdK9IVFSUbmMmRr4H fwhHjBgIjWXoa9i4jaAnY59nXbEWgAqHgwpXWxlHQrM4j6gmgavGjRXndINoi54lSvDYiebofVEY L/jINY3mTRuliNv3cNVYjKjzOKxDdO2FbRQYkTpojQQ5fQlYgK7XR8YGikH0XrYCJGc3gUlHhTQY bppGjuOY/mf6pITEJpDeBzYl2Wmk1n3ihkjAt4QmSZFxhIsEQ2ChVH47GXYNaZMgiZMwyZPEEXcp mX9dZIK3FmggRjo4xmWAeIFXhocGyJTyBpA9VWxS1HsxSH9WCZWUFm3s6Fkn2Y7MNpNnyEgDOJGA 85OeiJW61IPuJ1lriWI+qWs+mGaacpYLiUVAeG9PyWy9+JbREn5q6EGf4l6FOH+QSUz54oLGhJeL WYobV2wbeXaJV5WiqJOCpnkHqXfqSFjTSIscOZS99FgWCVfuNXj3g5qP2YcxiZGtKYyl1JivV4sB mZHriGmzKZFOqW6BuXegaYfuNpAmh1JwmIa1aZg9xnK//tlsNbhjjLiKDJdwBNk2vzVGY7lZrmhw S8ab5EaWxZlY8jie5Bld4pmQREmNYUmG73eDMqWcHfieGlmQ+yaO5bg7lqhSa5aaHueeUiguzDmN 0wabldWb4DmdqeeIYURrYMefVzeDd5WIBYp8/nmTI8qgGnSgg5SgmFlvkHeh+JlXy5KOqOiB6Il9 Frp1XPdJ6xmRHJo3VVWYjLd2H0qjmYOi37Kj7TecJkoqSNU2Rhqku/kW2hlOQ2iPioabPiqZcPek zTGlOoqJkRelS7qdaPeiUvqcMlOaeUimdcWlqZdXr5lZvOdwzcmmPOameZmekwmcYTpsdWqnsjhh eOqE/saYogT4p4CqU4PKjYKqehCFp4jqIeeYqFQVFxKKNJ4XdVu5ptu0n5Tqo84YmZLXgV9neLIZ op/qTlYnqtikoiXalFRFqKmaShtkMw2IkCLYeXw6q6o6ihE6oPBSqqbqbJvEq5WaoyTHntDJkoEF cblqrF2anrT2dOY2S3d5mdDqQ352nCazp1laMcKarc4JpiGpbgAaXGYqroAkoHbJKdf6Tyv5WKuj hXqqrnF6qUElWywUUURlRW8UjaEXmMoolN1Hk+oqn8K5ihqYLrPmoUP1XWQnr+lKnMszPitysLRF rjB5ahP4aIWTGDs0dUVJV/UDoZA4NzKEsXnic1ao/kKMxi3iQ3uGM2czW6GnGK/MiqrdWUIxVJd2 CiBdqYl1dJDD1SMvQq/K6jsZlrB+6q03KmEXq5SRKnJ8KXvLmrOsZbTdMpOssrT1GbDPCpEsakNR q7L4dHFKhKUZ24dQ8k0d67UUW4800402NoesWHtXabYWy5VAp7bPuGH4o4G4YmHEmXGC4rZAZkZC KiabiVy/qI57m2vuiK3yw3zA4VytBkeMwp2xSarEVLVva7cBarYh1pdYq6EsiLWOqZAf9kiLM7Vt GLlW256jO3ZLUpFRKbN7o2CUA7tPyLMJBVkp57sEpxxAy0AEO7tsKYknh6Gki6uFiqAqhrhKG7G7 /ju9t5uItfq8CNu5/XiCc+m6y/tBhKi62pt99nqvvjpZ4CtCIAKDi4Un1sm5sKqo23usLseYdUd1 gWaw0Nu915dOZISsECui8rSvrqMXBBy9rJuyTEu5rkJxvrhUaqq8z7VZYTtugXu2DnRgVcipx2i+ eiTBrumlzLK48Cm7ZEuiTEW9JMjCyIjAaSnCTku8SkXCNbqo5Yu21It+IExbMet/W/Sn5/mfIeyz v1tz0cnD4+vASByN4MJfG/K4BSx46vu0s8pBxQh8Kly2TzyC4Ru6iFqP+EWkegth8OabjYNgy6g8 UtzDIWOfHYrCz7u+9KnGDMPGxrnB5CupOguL/hPLqzo8ttAHx5sXeGvSRgrsvP9jV6HKvVv3pghc sfB7nxlccXSVkpCcSmiqN4NcTicjo5v8pex6MCR7fQA8yhGsviY8wDeLFI/2IKp8pBpbvK9KncdL lbN8bp0cqEqcyrlpZz+8y8E8yQpqzIVcypHotAa8tzZMzLt5mAvYxPCUVfGJxcXcwMMMzQa8ohw8 MURrs7vKt/unzdvMzTHszcIMzpf8t46IF3zkt+unws8MzXFnmKALxy+bz9BjmvLczPlcz7t8z9XM z3tWuwDnNXPrZO0sosB7rug8oUjaVn1HwxiolRgZztm8lxE9O0mqvGuMNpzUYHfZs2Wn0c63/rYd XYlveJSVDKKcu0h+u9AC06IprdIr3XS1PM5AqpJpYtM3XaACndPnDMwczXkeC3pFeYjjOdRE3bys XK+S50x1e9QerUnM/NQSzcAP/L/ThspYndVaHdRc3dXLDK5T21ThOtbiuG41rYysWdZ/OzlSDbAn DdEZ+smfJLvfSM15W79n7dW0U9eyqphjWsKBDIFCm6ZHKVEVjdBXW9hPRsQ1fdJUbGo428ivHFcv SHzlx1zGSc69C0YEDcgme8uE/Y/PJb+nfMAOCpp7dIAfzJPGdxi/d3N6vNVPldliasahvdl0rMx5 HKcJydc+fXvUyWbKlGmlvdPA/ZDB/WLz/onX8fWaxW3Wyj3EuevArkquUVjBhg3YOA04NapBVw3B ehnYdnTN5UhBslyCNfvbQWtFyym+G219c7yuLdvQ2Py52D3aGCaGSQ3Dw8WZKUw0txmOup1+LZbc 1QjU8kfa3nZZ3Nmj2OtoSauUetPLVYzZZWxYGll9nsnZfYO7jUbPTvywwS2VFIOSKb7hzp3OHAWj C+7ZBH24TTzg0fzBFu6Vxrvdo9O4Zqe5o5faZRrViS3OM/7h5WPbWJkiBgZCrwvk/u0nzf09Fu1O HN7hLczk69TcOohLB1dTV853oHTlQvzF2WjkECzZwm3eXq57YP40rXe92MuL8z2YUNvD/kO9wOvN xPQbR0ut35mY3y+H5otd5fR950fy5yAYYO2Cr5vIyCre20vTxRpO4ONd6NnyKIhO5Z+YSGROiHOu WqUefjiM2HoNPN8dZnlO1p0e5zh3tz6estqWf3y3Is3VWEQGS1neio+cwxiK5XhHd67axpot6wQa 4Mm5r8MkX0o96OJ9xL/uosSFgu8IY3Tzz96r7DpVqPMasf/908v+xwuFvroL6Ewo1uPs7a4p40vc 7Vtc7tReiXLaZc/u2CWK7ENq6EoukV+rnOx9yz4Si/PY6sbO7m6e7O4+1+0tnXAJ8W3upJQ+4q/u 7N/G7fIOQQwo8b7J5ZN9r/197vcN/u9bHk0dv/EqJ/Lm3iTaGIgqP+2bbnc0DvKHTd6v7doOu0tI johvzvE17zcvn428veYVz/NyPN2nGfS3mOjoetkq26TcA3gpb+1sztY7D+dGbZf+5dRYn/WCjdcw X/Vf78k2L/Zln/aE9e8Er/ZuL/Rrb9VvP/dNf/Ytb2rAm2oPGTRgNeRJjvcbyY85KuAIbnqFDztG I/ObTNhj02FImJjh5bbWjOc0FTHvOj18bED3QvlYz/iX49AJeCiKec84VUD4p3EmpCq0PdvVq/hn 7PERrrChT7PHjbfdjJzkHlKpn+bcB7g4nvZGTj9dLzRvSWe6Dr9XD+njI4ZUktQo/ll6O5/8icrm wr975hckNE046SzguV/dx8tgFGXfmp6+it3uxHj9/Rze1IrTOq7Luc/sJdWw8e31xCbTMW+zkabj lT/+tKuSUg4Q/wQOBDBQYMGDBhH+W8hQIcGHEAEsLDgRokGMGTVu5NjR40eQIUWOJFnS5EeLJ0s2 VMmRZcuRE1/KbCjzYEqHBHHSTGnzJsyWNBXu7ElUqEujDHnqPHqTYsWlDn1CfQnU6lWsWbVuVcrV pdeLYMWOJVvW7Fm0adWujVkVLM63bOXOpVvX7l28eX+idau1r17AgQUPJlzYK1yzf7MqNtzY8WPI keUiTnyWsWTMmTVv5pyRctnP/nE7jyZd2rTg0GQv+z3d2vVr2FtTq1a7mivPpxWdNk3I1KdSlr9t UnW6uyjvjUmBT7V49DfS57yZM13+HHjX4rG1k5xNu/Zcir2xIySPMXjEnObD9kSfPnzYr/Dfp6e/ +n3N9vhzllcvnv52AJECzzbWJvOsvf8S0k+8qubbLz+NzutoJwT9W9Cj8GaSz8EE1+MwQBD3MnBE tvgTsT/1LvywPgktdLHCFF9cULcOk5PoQBdN9A9FGgkMkbTuQKPLR7+sw67C+aLacUmjoosORRul Ggqx8nTEcCoc61MQSh63/DHAIFUjUjbwltxxpqZmjDBKFV+Eb00pLzqvypi4/nQvIhrfRI/OL7UL 07shA4WyxTPdbFA+NzskFM4ja+yxxij1dPApO9fTqU/Y/gRUUBKl4nBRHSdl8D4I5YRx0FITBUlD RFvV09JXMeVM0005XQs3BalUrro0o8KSOfKOA9ZIzyg0jlACjZTuuA0bbM5YWUejdaxpo7X2Wmwb q1asMbP19ltwS9Sr23DLNfdcdNNVd11223X3XXjjlXdeeuu1915889U3M2BFFM5JqoJFDldezWvO N00JzpLXRYsVOLvhenVOyesCPtTgHimWNTglFS74WOoe7oviZYMNGVeCPQ5pOEYZ/rjA/9R0VUtF YWXwUptXhVE/Kxlr0+Yk/lul9E4uhybaWlZvnjlo9/KcjdSl8ezNRFFFMrpU/lYEysojua566pbf zFNVq4tO8FDFLnwQ6JZDVfrGQs3cOOb+WpQZQrW7TLqmTyvNOz5Jz650sZqzzjLDt6UWG9K1y0Zw bDMhX3jOLsF2lcrFtW48WsgltHtPCznW2UvQ4QY1Udt8Flxx2T7jM+kHqY7wyYf4zs1HZBevPDmN OwfYdsqarHxkYn/82nIPfy/W4CvZ2z32on9PdsrSxS5+aya9hP3Rv92uHVXGwwa+YY3BRxJOnuNz y1f1ryW57uHhD5tx8oWmm/XNR7feOs8P0328yUHvVReD28zqVD1VNcyAy6vzW864RrfPYSuCpqpe 3vDzt/85EGcZzBlKzJdAaqXqaAy8H/I2dzwMtu1xKzwVzc4XwKMZi4DI+1mfJjjCu2mJQtCaXw1d GDcackeEPyTi1oR1ndC5Lj++YuLApJc6kinnX7mZ3cOoc0WEWbErxEEiFvuVxY0lRWEe+9fHuMhF 3qEJLl9k0e2YdSKUAOyKU7zevux4RzzmUY975GMf/fhHQAZSkIMkZCENeUhEJlKRi2RkIx35SEhG UpKTpGQlLXlJTGZSk5sMUUAAADs= ------=_NextPart_000_0000_01C71122.93BDACB0 content-type: image/gif content-transfer-encoding: base64 content-location: file:///C:/Documents%20and%20Settings/Klaus%20Holtz/My%20Documents/My%20Web%20Sites/mysite3/THEORY/info5.gif R0lGODdhJgLBAPcAAAAAAIAAAACAAICAAAAAgIAAgACAgMDAwMDcwKbK8IAAAACAAICAAAAAgIAA gACAgICAgICA/4AgYP//wKDg4GAAgP+AgACAwMDA/wDP/2n//+D/4N2cs7OP7ipv+T+4zUiENpWM QY5eQqBiemJPrB0vvihmdgBFAEU+AWooE4U5akoyhQQEBAgICAwMDBERERYWFhwcHCIiIikpKTAw MF9fX1VVVU1NTUJCQjk5OQAHAA0AALeZgYSZtL29kH9/YGBgfwAOABsAACgAAAgJKwAdADkAAAAA mwAlAEkAABEROwAvAF0AABcXRQA6AEkRERwcUwAW/ysA/yEhbFkUFABRAEcaahkyZwBhAAAx/2EA /1MgexZDZy4u4iZZFlFGBGguSQdSj2oYuJAjFQBT/6MA/2pKEnUzbEpBmjdlC6QsFYMfsU4s/yBR tghkkm9WC1lDrTZyErAzFwChAHdfH4lHcbBDHLctfQCGlXpuIyafAHOpAQAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAADKAKxbASAdwpRScCSqTAqUiTZue0R1kP8AqABx/98A/1aRSjRI+MwyguRBcGjK ATa8QgCa/5Yit4V9MyW3jDZa7Vz/AP9IACKbokLPTcJYUiDTlaUk4HNWtampANBvPGefWInPC/+s AKcu/uJZf0zcZ/8Y/zp9/7HQGMf/AP/iAN+aPVaBn8ZDuq9xiziiydFTzv+aZUbK2/9N/8jpakze 4P+Y/9/AgunspfX2zf/Q/7GsWmORriJMZY1OP1BwcND////n/2lpaXd3d4aGhpaWlp2dnaSkpLKy ssvLy9fX193d3ePj4+rq6vHx8fj4+LLBZoC/eMbw8LKk//+z/9GOo8PcN6CeVHaucHiewYNkv6SD 09E/Mv99AER4IyRfYA4OLL4AAP8fADE5ANmFPgJ3hbDYgVYhHQAAMIjIs6B5AMDAwOpwgVHxaf// gJF0zf98/6L////78KCgpICAgP8AAAD/AP//AAAA//8A/wD//////ywAAAAAJgLBAAAI/gD/CRxI sKDBgwgTKlzIsKHDhxAjSpxIsaLFixgzatzIsaPHjyBDihxJsqTJkyhTqlzJsqXLlzBjypxJs6bN mzhz6tzJs6fPn0CDCh1KtKjRo0iTKl3KtKnTp1CjSp1KtarVq1izat3KtavXr2DDih1LtqzZs2jT ql3Ltq3bt3Djyp1Lt67du3jz6t3Lt6/fv4ADCx5MuLDhw4gTK17MuLHjx5AjS55MubLly5gza97M ubPnz6BDix5NurTp06hTq17NurXr17Bjy55Nu7bt27hz8wXA+x8Agrx7DwwefDjx3wWPCx8O/Lhx 5M2J654eFvpy6L6Baxdo3SDy78nD/jPnjrB7TeW/lT8v7tt58uXce6uPL937+/rtv8M/Lz869vn5 sQdRcfoVKN56+5X330HuAUifde4d+OB18u3HHnr50VcUdtl1uJ2H5IGYEIfkLTieiCfeBF6HJnpo HoowvujihymGaGOLNkGYIo4vkriQjjaiSCKODIbo441ByqikfRL2WKSRUNYY1JFD1rjikUEyx6OU UeLEYYs+YpklkicSmV6T24n50pdjysgiRWwmGV6cY4q34pNvjlcliGzSOWOde6Z5FJVMiligmIFC KCCMAerkp3kmEloomDRm96h2aq5Z3qR21qkQlmYaWimTZ46op6CdWilqmVwGqiVS/pJ+GOedeCK4 I5d5enkffPg1WmuZvUZI5oRzOrrpgbMu2hCowirZ64ilJuhrgG6y+Kyzi1YrpH2Z8jQrp6j+KGGu fxZqbLhdkjtquX8OGSygO8XKKqrdriunracOqO62fJJKY7OytmourkIteeu8nvpJrsH9qojsumqG eTC7dFZbb0vfPoyuvgEjPOzFVw5c7p6UdvzxrzEupey01Cq6soYwtywggA6ep/G9RDbMMMMnj+vw vUBTzHHDw/bMEKQi87zzzSNvnDJ1Fc1MYcsTulsfgRFKnaGGWsec08rBZi2sQ9kKh7WiW7fnKZ9W R2fcegiizbXY6TXIq4XSQq33/t589+3334D3jd7ghBdOOFmGJ6644oGHdrFGj1MVuUeTN+5Y5RZh DpXmkFt+mbJf9gd33P5d11W23ol+YYKhP0uw54wp/XbQWxaNlZOy8ixlzowOVbbcXkvN+Y8Vak7l 3cMftrTTbmqbPFG4Z6m7pWiKDH2+1Hd5Z6m9Y6Qt2RJr39nySaua7vO+ewzqqNwH/TpQSJeI/fbs dvQ92XjSzxn5Pjc/+tpTiV7vaoatvKEPJvHLXq7odz/vvU9c4GrgZPjHPvMtjCsCXJ+V3jVApSSQ e/rrjgQv0r6JhG+BqVucCld4uAMWbGIZM5ntACiVDDItVDOk4U8+2K751c9+/t1bVv5yqCsMTkxo R6ydC+HnMSS6T4D2eqH4eDijEWYuiEcb4g+/5hXQ3SeF/2lW1k7HqxS6bWvCS5vodOgTHlJRhA8k oQMjGEeZLDEpd4SgWPI4NFgh7zlnjNnLgJi5MqbtkN4yCx9NNZZF4g92NiuLI1HWRU1BkiaTTB9K MslEl3Dykg/5ZCdPIsqelJKSoGTJKU2ZklWeC2OprI4Z0WagQZqwlYizZCwpJ63fqRJZ24NiH6W3 RpY5iGaVjKJJXOk3/4kPl1uqWB1TZy4gEe2awryKhXTpxzAC75CoC6Qa8wZGcl5RnJ5kpBWBGLoo rhNlzkwU06b5FIWphJmU/vPhG/VUMjiykXckLBk3U4VFQgqJVvMc5hb5taoLns5nuEzaGgk0t0+6 0YdNm5c/C6rMcxJxk4yk2OA4MrZXHYuj8MSVPHNnSCN2lCQnvFEwmyiSiz7ThlOk5yIFCsuTjvCA MVSg0SSy0vJJD3sorSdEQVqsW0mTjZ3L6U3Vp89pOpKnVXNdRjT4UXrq0UhU7OpXbYfQ+jkvmV7N Z1Nnt7Z3zhGF+yyaTTk6SaxGlGRiTeqnKHQmXwLSlmMlWu1o57TbLZWUCTVbSEU51xBSVaoFzaRd SYmfuqUxi03xIjqNKcbKkhGgiC3dprjK2KpOVa6mxaI9typDlVnllPj8/llPj9qvp+o1oN4EpG4F +cfdAjZuQA2eORGpNi7GRKusZeoeuZksjOZVM9PLkR0xtBHY5jKdvO1rbyuaGrfOdk2FS25ol4vd S4JnuB95XGxLd7ioRZS83/UckLzZXjmGcrorJOp7ZVle2GXzpB4VIi9ZSGD0vtSg/I1v4yYLYPfe 97gqPNo23UZf9loYqq38LW7ZpzqzsW5XBjShhk8TMgw/T734rS9mscnWwQo2rSUZKUmbaE3euSqm oaSuJ7fr1+B9cbcoHbGD31Mkuin0lg9GoIzh1NqyOvNp/FFxVK/pRJMeEcaG+yVkHQvWjkEKxSH5 8A6PvJKaMZmmbbIg/q3Wq6DEVffK/+JW9Q67Vzcrl4EYxSkKMXzbiKz2lQJW8oxbuzs1hzPKER4y lS+sPuRWjsAxTi2Xnai/pNYVU3n8pqJXnNmEppm2Dh1QgUfNwjO7L86iRSqISb24kTSWTHquNJjD 3Bw+L6vSmw4sHsGlarMKLEeJ3jQOWWwyeeW41cuUNKwfu2dLu7rJUh6t9fxMZqOsWYYuVqKKspzc J2f7ZsYGX3jvqeweovaZzq4poa/oKlMH2oMcBPKDGE01Wy9zyVv17I9htjpaTg20dTbznZEUV0oP 9VN97qN3Ed7QrAJ21gHc70wEDpL1PrqkquTxdseZW98a2MBI7vFb/v88bT6zOdnKxaQrLS7iwCHz kQ3W9VurRrqaBzjSjRzoLhm+L1SaPL3VTOKBeY5y+Gp55wM7VLcQVW33Kgyh0W060HPeX6SrNeZE 79zTQQ3wd+Pc6GW2OrVh7vOyI5nKUMfXzWEKOfX0G+TkrrrYZV52iOf70Oky97HHRdH/1Zvm9j5W iZ/r6pYefe6cpnu6p0xWrltV6PGBPNAWPlYGj5fyg8ajxjvuY3RyDrkzZ2vFk2x2OusX25J/dJM/ 3VAch97ycU/4m3e95WV3GWFwD7r92j16spe89HuvqPB6rNjEf6+otg981itsx6HTmvZwzTOzux5S GFe+6Ik3vfZ9/mfb/hna8OI9NXid33tYlTusPMUc72deN7b7fvvWZyWaGfpU6o8dzseNP+OtfX7n pt/dykd0xcderHZ/wKd/8bJ69OdpS3Qp0hWA+7ch/Ydu/3d2CFgrmDd1WFd9nQZDqEZYvVZIciNk Ayd311N7esc/6vd77paBV3eAELhtEyZvnBU2+qYV8XZ4/IeC6Id6LaiBCxV2HKh4TGFdVKdgG7J5 4nRZwmU8JChgIodAWUeErpVyYMFyE5QWdhcVRgh2sScZT/haUleF45VgOoh47vd+myNxX4GFaDiE TTiFcqhUVtiGOveGoqdRBLOFa1iHaHWGePg+uAZ/MQhorcM2/sjzOzNzhRdoX4HIgoNIfieXXoNl OpKXUWTUiGuXhJwncnhHTBd3KCS1cS/Ia0o3OZPITgqIfFuUimHnguG3g9F3WtZCZPP3VQ3oXKUI bhG4awzGigkke76oibl2grO4Z8IEewx3VYRngXOGPq4YVb/IazIFfoYVhNgni3hGixazbuCjblYG hPSiidE4ct4IjEhVjkUHi6EngTzof7onjDv1Z+MmbpjGhO04KGNEW+gIL1sxH244KBPYbDcWf7l3 diVEhk5RiR9IbHn3j8QIgO54jJPWjbcIQcyIjWxYjB5UKWCjiIloiWgVkNpIcNI3Q8oIYDvVjGmY j3RYhoxo/oLQo4RAho9RaI8s43Q3CS34uIkS2YcwaYcyuWC6yJHOyIUb+VBDKV8XSXqOGHF+qJRI CEkLAo2zJzlJ6VJT6V9NQmoIVkNZCZFLSZSY8oBYWYdng4FG1pWVhTeg95NRyZRNGRhdeI7P+CuT BYuL+EUGQlytM0jqmBkpCRh1iWbtY1eHuW7B5ZEtBjHxSH7xQpNxWGiE+GNhmGQj5mQHuRsbeVYM uVYv5i+FNGfnVpYNGZHltY3IyE+2mIdDM48saRmFGY64l4MjY5vzxn7uVEEOWJmvpJoVSZmQWVha F5uYpCA02JFRmZDUomp/SUEPaWqgRUA74mjKSZHJZ5HD/kmbqqiRXxNT7Chb2cidMeRioeZ6jjid HslBXCWQ70iBRiViK9lswHWZtNZu4WmW4/lruohDFWh/ejRs5vlfgXl62JmCwmmQhahF+QlNc/Jf 7hmUu8NXWtNv9ekyHvaWZNd9Azp/BXp/wJmdCRqADUo6JYpy7USfzWmM+1kVmrWEGZpd47Roxhii CMp6Coqa4fKhWqSHAMSjcAiOXhiWUyKZ4MRXMBqKufmULrKZShZvkbid+ReXYrmVjyikIhqkiUSl OHiHV/psYMV5G7ilEiqVgPilbcSiX2eGQoimUvSmLaqVZ+qmXAqDkRSUoGM3NTlRbhejRNqSdNqm Tjlm/n54QrITeVFybV8oqIG6qF63Q6w2apQUJlCEn1LKpI7KiYEUhaWEccHnU05KpoMKF+FTlXnV m3UKpub3nk36XiuYMFEaoYZBqYYEQp63aoyaqZp0oFH3fMIIL7Eqi7P6JJHia3x3qT5Zpm9qo71a fskjMbEaqdIqrYmBnjXmfacZp9o6SsyKfzVlnxA1eEC6YWaErHKqqN92rDr6qH+apqwKoFfZjmH1 krYVqp/1dn5qKB85gvZqlNtKqLzqrb6KqVRzonfanuMKhl66rCaJbr9qoOekQQZ7nG31L27XTAtb MEa6kxXXryvaeh5LpsaWKQlLGCTZqBgDmNpXsoNx/rIoO5Xr55ux4bIvq6uKOoe6QbM1e2/c0pYR g7FjubNwukfFZIPDF7KAqqtC23xjmBXXmnrr+pVKu7RS2LTaRFA0tqDK+q+Q2okzmHF9CZfWKKfs aqYVxI9aq6pzOiXlVlz99Xk4hrTQNxfhFk/g2q5YWpJVJLBlaDwMZa4KKRd1az5lJZ65WqMNS6P7 5bd9QnION62QG7l4EVSN97CpmrcTaaNpGzWeilkpCq9Xu1mNWaXpWppFZKXcmrigW3595m2Am1l5 WWsYxDogiaFAqrNsi4Kb+4NEBTDiJ5QoyYKugbupu7e/m3LDw6GKBJolsrqoQbzws7FyCyf5ipCT 3XmErTKIkTo+GUu1TEuec9ka0Ou9zKVmUSuY3Uu+5et4u7s/6au+YLupHseyk/u+8Hu/o3q5+Kuf ejO++8u/UOO/prRx5ahjfmZOmuYW9EuX9iuysFbA0blix2OcMck3ApyA8Hm61itn3rm8gtPAGEyQ xiWdRUZEkXvCKOyVf3PBv3lwFEvCqTKx/8u1mIu4mAjABzyDwTrDIby2xQuhVYupF9u+PLymh7ur sjbC/qqiRSx/qBu9XquOdytmwLLATZzCBdbEWrzFXNzFXvzFYBzGYjzGZFzGZnzGlBEQADs= ------=_NextPart_000_0000_01C71122.93BDACB0 content-type: image/gif content-transfer-encoding: base64 content-location: file:///C:/Documents%20and%20Settings/Klaus%20Holtz/My%20Documents/My%20Web%20Sites/mysite3/THEORY/info6.gif R0lGODdhFAKOAPcAAAAAAIAAAACAAICAAAAAgIAAgACAgMDAwMDcwKbK8IAAAACAAICAAAAAgIAA gACAgICAgICA/4AgYP//wKDg4GAAgP+AgACAwMDA/wDP/2n//+D/4N2cs7OP7ipv+T+4zUiENpWM QY5eQqBiemJPrB0vvihmdgBFAEU+AWooE4U5akoyhQQEBAgICAwMDBERERYWFhwcHCIiIikpKTAw MF9fX1VVVU1NTUJCQjk5OQAHAA0AALeZgYSZtL29kH9/YGBgfwAOABsAACgAAAgJKwAdADkAAAAA mwAlAEkAABEROwAvAF0AABcXRQA6AEkRERwcUwAW/ysA/yEhbFkUFABRAEcaahkyZwBhAAAx/2EA /1MgexZDZy4u4iZZFlFGBGguSQdSj2oYuJAjFQBT/6MA/2pKEnUzbEpBmjdlC6QsFYMfsU4s/yBR tghkkm9WC1lDrTZyErAzFwChAHdfH4lHcbBDHLctfQCGlXpuIyafAHOpAQAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAADKAKxbASAdwpRScCSqTAqUiTZue0R1kP8AqABx/98A/1aRSjRI+MwyguRBcGjK ATa8QgCa/5Yit4V9MyW3jDZa7Vz/AP9IACKbokLPTcJYUiDTlaUk4HNWtampANBvPGefWInPC/+s AKcu/uJZf0zcZ/8Y/zp9/7HQGMf/AP/iAN+aPVaBn8ZDuq9xiziiydFTzv+aZUbK2/9N/8jpakze 4P+Y/9/AgunspfX2zf/Q/7GsWmORriJMZY1OP1BwcND////n/2lpaXd3d4aGhpaWlp2dnaSkpLKy ssvLy9fX193d3ePj4+rq6vHx8fj4+LLBZoC/eMbw8LKk//+z/9GOo8PcN6CeVHaucHiewYNkv6SD 09E/Mv99AER4IyRfYA4OLL4AAP8fADE5ANmFPgJ3hbDYgVYhHQAAMIjIs6B5AMDAwOpwgVHxaf// gJF0zf98/6L////78KCgpICAgP8AAAD/AP//AAAA//8A/wD//////ywAAAAAFAKOAAAI/gD/CRxI sKDBgwgTKlzIsKHDhxAjSpxIsaLFixgzatzIsaPHjyBDihxJsqTJkyhTqlzJsqXLlzBjypxJs6bN mzhz6tzJs6fPn0CDCh1KtKjRo0iTKl1aEoBBAFChPo3qlCDVqlapTo1aUCvTr2BPXsUq0OtArk7H hl371CdZtnDjWiSLle7Zs1XrymX7NmffvYD/qc0qlbBgwVfdWi3b9vBdx4HX/o1MWaddxo8hF9bb 83Lfum/zVmY6ebTpmZ4bpw6aGPPirHgLnz5aerZtlqtfI+7qmqfozMAf/74tWWrawboRmx1bm7jc 1pCDyzbrO/Pn3sOdhwXtOHvv6L+9/msPnP266u+WgV/ejXn9Qubw48ufT7++fehF3U9Omzy0w+YX 3SfggAQWaKCAGHEWXXuvAVgTV2WBJtthECLHkINKYbjThMHdpR9vDWk4kYjjdSiRguDhJVxyJaLH 11Ic/qUgfi4eRGJEN5aYo2ERxmZXhTS+RB1MOw5VJEoauucjZzL+J9KRxCUJIVzw0eQgchaiFaFX QB6XV2sWIgSlSbV9COKK6GE4pkJSJhbmd9NtZtyWXRlnp5tqYRmkje8xx5d8Vjpp3aA9uobioYTW aKJNZSJ6JoNpxpQkpEoumKJm3MFGKWPhGdrfhWIC+pV9gYaom3mJfoodn4u2WqqY/srxSOeWU8a5 ZoAnPupfaEyeqtmvLMIJq6tw1pfhgK++5yuwrJp5qbPDbhhlrqp6yNtwHFJorbCNoakeqAQideC4 5PoZInXLJYTutXVOuGe7VdoYX6jzypvthqJuZW6woBIW47faasrptgs26S+8V5arMLkB3scRlGpy iyKISvLX4MX9tnjTjf5Zip9oBncnMrePjuytsrTSlzFQKm8EsaCXsjixsB2fCeCtGnvEsc0yW1fe itiWTDLPKy+5r7pG1avRy6au6q2ZqAI8c7c54/gktanW7NmMAzPrYs3MXslmvGwm9eZcuF4Yp3JT xmr0rNKBuTbSVWP9EYm1rp23/pdA29renbG1+zbc9/Krb+GG54ezpVYDtvhtjwtdN6sPIU4s5Gnv Ffndsqb8ZeCBS6hVl1TrPHnRSx++n79tM64Y2o53JjnXALM38s+ld7R5ZJuHzPhqYLPW8MLEF382 o7PLm6rHuqpk/PPQQ8956bvmHrBQu58eUvCQdj11yZVqvx3lNFtvMfbi/8Q92xcf+i/r2acvk+8x c//9uUOqF2bE8qN/HtDSoV7i+icu8oFnSJ/JH0TCBzzJGZCA1TGfioY2tPhBsCX0m1jFHNi0AEaN YGW74Ou6Z7KYMa9aIswQ+TYYwAGizIMSFFsK8cUluWmwdW5zGw4tOEOQJBBd/n7L4d4oAjYNdutm PbwgD5OouxFiB4GHU1v0pkjFKiqsg9Jiovqc6DUTFkyLRMJiesAouzI6rXYUJGNKZBhBNY6Rhm0b ohCD6MY1irGOSsSjHqMVwj3m0Y+AdB0HA1m3JRLSNoY8JGV+WKgcKvJ0eKNRkCRJyR0q8JHOW6EL MamdRHLyOSu83yc7Ocok+s5iniyl/1QpwlN2bSVA/JKXLmmYWe4QfkZLJXnIM6ejPXE5x6ujK69n RzR2SpCGQmWkVHVMNeoSNypqZgs3hUxnarKRSDpiz7AYPmKyEI/PhCUKFwOtagrzmoq6GtHSqC76 gfCb53ScO1/pxXAiciuE/hPnOr+4wPXNKH/2jB3v/OmzWz4yoJXT5jQXyC/oiMecM0QoNE3k0K9h UqJ3dNTlznjCp1kPjBhtyjj5ObOQLtKJ5VSTRltY0j2adCTTCVbHWnpQ9VkyRpbzkLt2Krq5OdKa mnNkHG2VN9Ad8qWsTFpSW7nUHiK1qVyEav+eKtUsmhKnNTRqrHoKuJ/uMoWtGwwdt5pLHeb0RQRc nUJNmKnbcfSk4lPrOp0VU7duVDIQlGtDadehIp6GqjjRK0X5ukw0wnWqfDyZMtNZwb+mT7AkXKw5 /WoawFp1dnF8JV3eN7jRWBZ5D2QfAH8JvncFNa2JjaxmhUZZzz42tSUE/tlCG1vZvMI2ZHXta/Nq G9fbUi63hX0r72x7zdywVYB3HV9vi4sxjn6wm6dFrOo8l8+DCW5WOBSo9mRkQ9IVK1tyzG505fdZ xy4XquUNbFXTi5r1Ehe9r5Uqe0GbEVrGd7suEy9535sgX94XkvkN5uTm+6AAk+28AF5avhD83+HN h8GFNPCCI4xaBSOoagQOo4UdhuG8WvHBGsuwkD48YVJWuL4cpjB+N9wyFe+Xxf518YAlHOMOn7i/ NZ4xhCuitBcPGMRENO2KQwzkEQk5wT8ucT+ZKGIkFflETVbvdntcyiiLhcqjtLLzsPxJLTeFy5z0 spMPXOVCgvmijyVz/paTLGBFinnMR3Yzm+1b07jmGM1mbrOc7azno85Zv3jmM6ADjeE471nQZz10 oelcZ0S/eVSQNLSfyTtoQi/60ZCmdKIVfWn5TrXSjc4zpkmT1k1POtKjhlGp3TtlVlPa1T+GtZ1l XWhaV3XHt87ZqEnM6177GtR2M9Kvh01s6ZlOhTBDHd3uuGxlw/bZoYU27PKT0GS/0NnRzvZHtT3I bhNR1dbuI7a3TW5vJ3dS4a4vbarNbGmbm7GbPDe7x81jcLeb2/GGt7zT7e58Q5RW9G5jwPd973L7 G90Ff7doxb1ufv92kvDKnZbiti6HG5zg9IqzmBGesTfxCrz4rNjH/uZ97ZIzXHVn3TXJxQ1PeGJK poY9ebNNPvNcMnrjK292y7fpM5gLt+ZA7/einozvLVp8tySLWms7qnCO0zyxpHq60ROOQq1Rbekf xLjUL/5donPdjAN37j6Fey/zrE/o+k47hZC1dbC3vepj96LtfH72oqt9UuF6+9TDfkCDMn0/9gtr ztH+b7YZr+FU15/flR4q1gr+6AevXPTszXfjLk/uI/WO02VOeK4qOdOJdxpur87y/2md83bHVNS1 W/llud6w5XTd5oOeeu5+ntSQrx2qGA93B86+8zlv8Xj1fkbeUxC4xv997QfvOWArNffFJ/1skS/9 wiv/6426M1iu/m8znxbM75ziLHVPT3vs48/51Ia+6ifYdQE+vrPWZ775pYj+Aqqf+5G/v/ybjqNU A1//ABh68ReA/+F/xDeAAoh/5Pd/AjhcBHiACnh3+5d/NpaAE7iAy/eAqDcbEYiAraeB5cd/olZs JFiCBUJkJpiCKih8udaCLviCMBiDMjiDNFiDNniDOJiDoYNdIAd/g4VKedJdZTcnZsU3hmdLWNUl YUWEHgaEsnR1e2J73RGENaRAE0c6SngnCCSEH1d4FUh9yBVcyURNhKU1GrUei8V7Z2iAZMJ+DIR0 ulUox1SG3ad7viJNoqd7bJhNr3cdKVVP4INO0Zc10GZ57IRk/p6iW3/ocgS1UGdoOKVhiPzEVHzk h+N0PtU0U5mXfM2lKA2UXDKWecdFTi6kiTHXg4lIQoL0iWqHa0c4V1CUilA3VPrFU7ViXa3iUJml Va64PwXFWaL0fbYYiZuVLj7CQbroLp1zY3BoeCeTMsRiiqr4W314iLH1VnsIUzqHMM/ojIwljSUU LWjYiauYh7PFjCsljSwUPI34d3Y4dMs2NXUnY+lIMbB3fMkTXI8Ij+JIUQr3hb3nUdNEU1LTV3T4 c+UoiyvlY9CijjxHkOboMQeJeQnpNQupRDzFg1wVTXR0U4NTVDolIXTiXctIOL5UcWlWjBrZWV01 R4w0fiAZQH4iuVVy8pLXdYRJ6G86uJM82ZM++ZNAGZRCOZREWZRGeZRImZRKuZRM2ZRO+ZRQGZVS OZVUWZVWeZVYmZWRERAAOw== ------=_NextPart_000_0000_01C71122.93BDACB0 content-type: image/gif content-transfer-encoding: base64 content-location: file:///C:/Documents%20and%20Settings/Klaus%20Holtz/My%20Documents/My%20Web%20Sites/mysite3/THEORY/info7.gif R0lGODdhIALoAPcAAAAAAIAAAACAAICAAAAAgIAAgACAgMDAwMDcwKbK8IAAAACAAICAAAAAgIAA gACAgICAgICA/4AgYP//wKDg4GAAgP+AgACAwMDA/wDP/2n//+D/4N2cs7OP7ipv+T+4zUiENpWM QY5eQqBiemJPrB0vvihmdgBFAEU+AWooE4U5akoyhQQEBAgICAwMDBERERYWFhwcHCIiIikpKTAw MF9fX1VVVU1NTUJCQjk5OQAHAA0AALeZgYSZtL29kH9/YGBgfwAOABsAACgAAAgJKwAdADkAAAAA mwAlAEkAABEROwAvAF0AABcXRQA6AEkRERwcUwAW/ysA/yEhbFkUFABRAEcaahkyZwBhAAAx/2EA /1MgexZDZy4u4iZZFlFGBGguSQdSj2oYuJAjFQBT/6MA/2pKEnUzbEpBmjdlC6QsFYMfsU4s/yBR tghkkm9WC1lDrTZyErAzFwChAHdfH4lHcbBDHLctfQCGlXpuIyafAHOpAQAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAADKAKxbASAdwpRScCSqTAqUiTZue0R1kP8AqABx/98A/1aRSjRI+MwyguRBcGjK ATa8QgCa/5Yit4V9MyW3jDZa7Vz/AP9IACKbokLPTcJYUiDTlaUk4HNWtampANBvPGefWInPC/+s AKcu/uJZf0zcZ/8Y/zp9/7HQGMf/AP/iAN+aPVaBn8ZDuq9xiziiydFTzv+aZUbK2/9N/8jpakze 4P+Y/9/AgunspfX2zf/Q/7GsWmORriJMZY1OP1BwcND////n/2lpaXd3d4aGhpaWlp2dnaSkpLKy ssvLy9fX193d3ePj4+rq6vHx8fj4+LLBZoC/eMbw8LKk//+z/9GOo8PcN6CeVHaucHiewYNkv6SD 09E/Mv99AER4IyRfYA4OLL4AAP8fADE5ANmFPgJ3hbDYgVYhHQAAMIjIs6B5AMDAwOpwgVHxaf// gJF0zf98/6L////78KCgpICAgP8AAAD/AP//AAAA//8A/wD//////ywAAAAAIALoAAAI/gD/CRxI sKDBgwgTKlzIsKHDhxAjSpxIsaLFixgzatzIsaPHjyBDihxJsqTJkyhTqlzJsqXLlzBjypxJs6bN mzhz6tzJs6fPn0CDCh1KtKjRo0iTKl3KtKnTp1CjSp1KtarVq1izahUIoOs/AF+7ih0odmzYsifB EizrlSvbtW/dopX7la5asnPxtj3rVu9ZtGzvBkY49i7ZgoHBDt7KuHFLwWoFrz3MdXJJw5Up19W8 OXPnyHg/Y/7MWXLpzKBJn0YcmrXl1KYdy07rVXHcv3uJGk5tcHRnzyEVu6aMObZqy5V39ya+/Pfe 4sOjQ0ee/OD02dhNQvasHLju6eCp/v8GKVw8b+7Iz+cOm9D0aPWbr6OWHt36eO/ns+vv2N33faTP maVXd5x9ZFx46eX1V2/F+ZbXe8CBJt+CcAlY2nMFjqfgfhxqtF5++QE432vmjXSbhvgxR6J99d0H oXPVlYjif/KdmOJ/HeZYUYAfWqjbis3V55+HwyGo2YFB1kVghi8S+GGL6AlZ4IRD6mglRbC5VqVP GMIY4ZciXeckdUgehmGX4oHp5Y2rHUfmlFJmeOWc7QEJ5Y9mLYZbZDZ6JGBht+nJV1+L2SbheoMi 9mdbf1Y4II+tVQjZpH7J5SOdmFrnXnppZurpp6AWZailPF4a6qmopqpqSokhuuqr/rDGypKpdMlq 6608tbqlUq4StiuuK/UKbKa0JtqUsAohC9SvWSk77JXORsrrRczuVG2YMkX7bI7aKjotRtfiFK6B MXW7LYfj3rnsRunS1C67MJl77n7vdvpTvSwOhS+1+/I7L6r9ereQrgQXbHDB5Or72Evy/qtfwHIS lm1wCgfrUsMOZwexveqqtHHHOX3skMgPYZwxdiQLDDJKKUesU8t1BgvzyVO1XOXMEc2Ms5+semwy zSibKPFMO+8ML22sGg10VDrnOzFJSmekdNTsLa1q0St7fNm6WyNt9apYT9mjofIKCrKPeTpLtUVh X7b210+1LSanKg9s781tGue0/rVQa/c23E5tiXZtGy6Hd19ZJ0njyJHqrfjLQkP9N+CBJ5u3y4jH DKdEE9atOG+Ccxl5mJNTXnlD260cepLpdo4jlOW97nlNNutcuumnM0Sl7ocXiSXsqLNYeOLxkmfi 7bjnbjnruuPon+OMd7rr8xxj7m7C5CGfvPKa0223i2zKvnzdq9cY/FIHp6/++upvP+f0Q4sfJam1 zp5sn5jHtWj034rqvpXKAlH12kS0rqHvKNr7n01ip6m0LWl89tNa35iSwB0pEF3WI9LjLmZA2o0N LgNEDeGkhr/vKao8Zrtg/+QntfiVa3S0IyD4vredfkHPhEeKTQVVqMEMttCF/gyDobik5SAT9gdc bGNeBHlYMeL9DoJBNF6utFS1/HnLicKbiIOaxMQVhpBazXuhFCFnOPh0j4j8wtiQcjO8Lgaldzzj HfvmuD7skZE1ZQKi957YOiMu0Y0hwyF/hPKxHZ7RTDKkz5HAaEFFshCQd4RiD994NL7NKI9OQxO+ HhhGEM7vkZB82vkq2cQk5mouFkIbBEflw/aUUJKs3FMo75UzUv7IlFJZ3SF7Nkuo9BGJSVGjIce4 yz9iq5fcg8gmj1WtnxFylF+cIDIPyLbWDdNDwiqWL+nIzW4aDIC1kSWDhKOnFBZGnFgsIAlLdk0/ tYox7SyZjlLnOHpeLkGf/mxlDDlSLG1Ok3+37NARuxcizw1UYvHUo6TK9s5//tB/8zTSFVVEUU9W FJG09GMtHerKBpHNdZUKaebY2UZB2kWcKeTVpG620hOusaW+68mv3pXQq91zRvniJE31mSZ7zq+m jCwoFS9JQ4G98nrnCxdQYZW6no7ySY2M6mqK6Es7KZF6LhQqSNWJOn+mc5q7y9rcTCbUTlp0RGhF 36bEVlFMdmlM0ZyVMo25VKZmcoPAq5o1yerKAUmqcg68FCvLaaPBBgpSN1HqRjkK0EXiVXr5jF67 ygemusJtXJz8KljjR9Vigk6LjHzT4hjbwZEdFZSMFaCcDjc3Y4p2rhM1/mNmSQtMitFWjoxio0Lp Z6zemraksBxUoU57W84JsbhxhChyuUrM5SY3oM6Nom2ja8dSUtdX5CSuSQd5XQN587vg9WpjzklY VJaQvH3yKHCL9FmpVre7tnwmt1qTR0xyp70yKutVH9pc+MaXkvO96GrjlFY87s20QwXtMf37XOvS S6LxkdFZI2zgx/LOrwqeLoMbeML65e+DE16siB310c5SEKYYDp9FN1TPoLqWpxnecJDgitMZOzW0 I75vZFELYNU+kLX4KRxmrQpbDctYc5oUq7R4bJ9fKhJvljWrVvdmvtFuTplrpetxj0xljH5xrPVi oDw1leDJRHlgWXaT/hlXXFkCyzOw3XoPiEe6Wy5DVmXwaw6jnshklqbYsSKC80pzOzzDmpdSHmaZ ja8sSTtvUEIWxvOOuype1RUYN0funW7N6ugDx6jGnqYwk0HN6TZ3GZK6GjMQD9VoliHaUo8a26LY SOgqCjpuN9IbkL3n5Bdvsa2MVmE/1bZL25BPjGqicIjSvCS3xpVLh3aUuvQ3Zyy/GppmIvSsRfq/ OMMReNBLYJId+blF6zjZo66qUc4MNmvulnotnhh+E6miF60JhRKeDbs1i8xlPjrXwY5i7Oz9OlSa WzUM3KpW9v1sh9rwcdTeDVTVOXAlZnGG9/40qccb3o53078QU3iO/ouXOYKvsUX1RWPDc4lA+BZy 5XXmIAG3avLXNhvmTGs5rjzecew9fIFogrW0O+zScb560Aw3snxtVUGbvTjUnb6fRxNtUBBr16V8 HaeSDnvoSlNTlHH0d9RVTWOC75HGRR61vfesJvuynLnkCvPYozruEDYV5yI/8EHzCmPoItt4O507 n9Fq9lPjPZ2/tuqan750sAdnsoIfvJcLz3fGb3y7JR/cpvsOYKQeL8aRTztxBLvd2WJ7yMvzcb6T OXpjS5PDQ68imUce+lKje3ZYtfzlgytgSeNcpjGrYRpJj+TVYzd96b7XtR90opQ2K3y517ObRdzr krfdTkm3/Sf1/vvUSOcwmtkn5s2TfcOrRHzJd4v2751fakPP+tbqpjLlVQ31OLm+1Y6p+0XLH6rw h39hxWZ6SVV/A8Z2xZR/osZFdPN/JKdcnhJAxbcjpjJTTWY9DMhdFVdmouFMGqNzoKJLGddIJvZu bkKAWwFpRsUm3IcpPNeCdXQqcKR/xnWABTZvJrhwjSN9ewQ2Hth/BGUschdwC+VhXnWB/PRaaoaE PrhuNhVzyYdOnIcldJRLQUchUMh+xNKDMHiAOFM01Vd7csQgVOd/r3Jy3sdfjyd5YBiBP8Vvjgcw C+Q27rWGGpVPZGhX+0QxQUiHYeR6uqaFTQh3PCN2fMhZKLh7/qIjK4bkdE+oe0yYKjpUgneoiJ4X dmhYFUboXT/4aXlnSbfigt/lc5OEiRHVK1nCiW74d9H1cmIzdUuWidrHcW3EddzWiA14XSHnSD/m hBT0MFiIS35HXT93NyrHi150gg31Xp3ncjgWMeGRTaQoi8nYYI3ncu7maUVkgGcIiFbxTY8HiuA4 jQzmbY2WjVb0dlhxMFvGNVw2bHlGjPxni5FkfgizYMGYaeIIRaoXQbD4ewiUapQmhr7ngIW4SupH dcboYN2Yj5xFJiNYjQWZkMD4dQuXGDOIcf6oihF5g7VFkelokTN4f1Z2jxuZio0lIjgodItlgA+5 jCUpkRPJ/o01E3tFFonbGEgviXncFUzPN1FpZ5MmKV05SYP/pZDbdDT41ZLsOJREOYokyXoSuG1j KJNMmZEnCZG4JoguWZUc2YxYeSx5+JRcGYW0l4hHGZZGOZZWiW1LCZaJRZVqSZai91tbV2g76ZaJ FY56eXVj2Tb8w5JdeXwMSZcllkm/qJNxmTF+KWW494TeeJGAuV/pxoGJGSuLiWYaqHFaBJAxpl9i 5piVmTyBx5aYqZKAtpmDiWBEpiQ32ZqhOXtA+H60tl5ZeZGkWVQuQptSB5LV9Gd2qIavaZvjtyYl 2I3GdS3vKGrTlzOpuVHLxmKc04//JIPKWZwp+WZlaX0h/iiXzcmYzkNutxmcxxlZqlWdPdl9c2lm 2oZYvkVSfAmba1ZlV8kXJSZxpXKfchZSCiIoh4k7h6iAOUSZcDmE2fkYvAla8JcoDtRVhqhsN3Uc JuZsD8pD/6mDjiWdBYqIMEkbulmJFkdq8Vma5vmHjmg1TQWgkwZPgoSc3ek3HfqGArmcIfqdp7iD acZEw5mE9HaeWoeQRAd7s/KeQimiKhYf+0krKLh27CWg/9Iof0Uq9ul1M4lN7GSgTMqcldYwuzhU EMYcy3ZlKyieU+Qv0SkzVyqBiElunTNvuqSNv1miYuqhGZqm3nWmPkmkzxZ0Nbd9N8anM7eWcQqA tjmH/pJDMmrkmrHkW8xXQ+mlp6nUfEIaqHE4p/OZhvwRMHYqqV+DnHcpOZ0ak5rqcJUKqtlzhE4Z qhyVnJ8KfKgqU4jCm7NYbVLqka6Zng00mAQjmPzZqqxKVMqppEyCgBs6qIH5erw6RMcWrA0aYjwK qESJocN6rAenhAWHcrIxf6sKp1spdZkncZUyXHxCOJC6qFgnm+Lqo9tyQzqlUNhqnM7aatDalGr6 oUCJcCo4YZF5igL4LOrap5WnrT14mfKYlobJJG53ad9ZgTFCoqvZpM7orw8blCg5NRJLsMzzpbIV o3ciQJJhgxASr54YfcS5f7Wac4JjdcS2Wig7idB0/ohe+qQbe3wVxppWGKwgay2mKGu3Gq40qW+a xmt1eHanSWkvOoY8G577xUV7R2cR+4wIO5DSOo9IVq8XNz7UiZucen1L1Ez4xF7EOJJfi2/VWrJR y0uc1rHNQ4GcMppnmHKxSK9wC2wxe6EOyq5l64nl6EPR0ll72LbJYYMJuaBBpkPnZ4WCRYtGWmjV drdoabc0ez+k6bJl+Y57Fo8Ay7ibqn0ZWCcgaGZRaIZWO3lIi7ldRFlAYrpVS3jAaT9jYrkDS7om 2lG1VovgyVuPyrYdU6Obi3+wW66x5hfhmqCAErzrebOXq49mWmeI257Firlld24QCrSAtq/H+F6L /qhovduyyrpjzztg74qTSne8HZm9YMdsKNq9adWJKBm+4uuVHlpet2uuwHtS3TWu+flnhaUhqVSx eJutcum/Gol2w5ll4cS0y7WPi7a0j5u+/Cu1Rfm9uBZWrUWenvu/buR2aDe0YjaCxtu+j9TBFiTB MRUl3sqsq/in+ge4e0q9eHmqW4hu5Wm/Git7qxggKKWghSt76lW8J/jAunpenFmukepq+6t4JJuw EEy+20qspAp6vfqmE7yuXva6SiyWo8vEqzuPbqu1X3ImrlvFVLiZ1Eio2oGawMt1txsahSJczAvG zdJMhurD6+jGqZVNmWrBfTeMdFzDB9pfZEyn/pC7x4Gai2aMZYL8W++3UHIWbfxZwH2cuS5MUhd2 yGRHfn1ar6Cjr6azMZjKu5RseKdJeWWypf5pqp+MVF96ezMKvQ3Mg5F8ylaayphmdAlyuGQ7LIQM yxSXgk87o6DrwZAox7pstrxsnXwbyq0ciOM7zDKHhOGWU99XpJD8yszMobSsn/bZYfuTqG1MM5xc zdX8zeA8zLk8zrBczuZMyeKczp+8zuwsrXvpgu9cmZYFwi+5yBuYuO15TsL1qoqszP07z+PrU7bm vQVt0Lf3gBkl0C10c1FcbqpMnPYMwJqHrjurLTssrhMYW7V2tGfMs5QSS0dKu/ysko1ctIqp/r8c E293NbMj+ymgi759+EcIsnjFnIB3d9AEHZmWPGNbvJ2iabAGtb1GN3GDBtDLap2A/LUytMHMJrpP 7SWyXIW+asTAbJmzOLYYCZ6BctXuuqJTTH+3KZ+aKbY0/Lib+6pdirZbraytNdFP/Ks4xdJ+CnBe TRWqmqJLHbGPMsIR1iAVjNZ3qtMJ3Na8LCaz6rCvQcA7WNcjC7gwbXuSO6qhttb7qZkknIDYeGmb Mn8oaq/bs22JmsYTOJu1Ir+IA9fuC495THsTPNf0tYCaLdUrAqxnvdb3ytC0atVf3HAia1+jbJNm zSejR1E5eswPmqO6/ZWznMZYmp5SyR6iXv1WOAwb5yrS/YxiCopd13yFQ7zcWoms4D2U9TzeTFne 5p2T8dyC6d3e7v3e8B3f8j3f9F3f9n3f+J3f+r3f/N3f/v3fyxKmv8u6d2weGo2QIltGGU2fgrvM AO4TAQEAOw== ------=_NextPart_000_0000_01C71122.93BDACB0 content-type: image/gif content-transfer-encoding: base64 content-location: file:///C:/Documents%20and%20Settings/Klaus%20Holtz/My%20Documents/My%20Web%20Sites/mysite3/THEORY/info8.gif R0lGODdhIAI5AfcAAAAAAIAAAACAAICAAAAAgIAAgACAgMDAwMDcwKbK8IAAAACAAICAAAAAgIAA gACAgICAgICA/4AgYP//wKDg4GAAgP+AgACAwMDA/wDP/2n//+D/4N2cs7OP7ipv+T+4zUiENpWM QY5eQqBiemJPrB0vvihmdgBFAEU+AWooE4U5akoyhQQEBAgICAwMDBERERYWFhwcHCIiIikpKTAw MF9fX1VVVU1NTUJCQjk5OQAHAA0AALeZgYSZtL29kH9/YGBgfwAOABsAACgAAAgJKwAdADkAAAAA mwAlAEkAABEROwAvAF0AABcXRQA6AEkRERwcUwAW/ysA/yEhbFkUFABRAEcaahkyZwBhAAAx/2EA /1MgexZDZy4u4iZZFlFGBGguSQdSj2oYuJAjFQBT/6MA/2pKEnUzbEpBmjdlC6QsFYMfsU4s/yBR tghkkm9WC1lDrTZyErAzFwChAHdfH4lHcbBDHLctfQCGlXpuIyafAHOpAQAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAADKAKxbASAdwpRScCSqTAqUiTZue0R1kP8AqABx/98A/1aRSjRI+MwyguRBcGjK ATa8QgCa/5Yit4V9MyW3jDZa7Vz/AP9IACKbokLPTcJYUiDTlaUk4HNWtampANBvPGefWInPC/+s AKcu/uJZf0zcZ/8Y/zp9/7HQGMf/AP/iAN+aPVaBn8ZDuq9xiziiydFTzv+aZUbK2/9N/8jpakze 4P+Y/9/AgunspfX2zf/Q/7GsWmORriJMZY1OP1BwcND////n/2lpaXd3d4aGhpaWlp2dnaSkpLKy ssvLy9fX193d3ePj4+rq6vHx8fj4+LLBZoC/eMbw8LKk//+z/9GOo8PcN6CeVHaucHiewYNkv6SD 09E/Mv99AER4IyRfYA4OLL4AAP8fADE5ANmFPgJ3hbDYgVYhHQAAMIjIs6B5AMDAwOpwgVHxaf// gJF0zf98/6L////78KCgpICAgP8AAAD/AP//AAAA//8A/wD//////ywAAAAAIAI5AQAI/gD/CRxI sKDBgwgTKlzIsKHDhxAjSpxIsaLFixgzatzIsaPHjyBDihxJsqTJkyhTqlzJsqXLlzBjypxJs6bN mzhz6tzJs6fPn0CDCh1KtKjRo0iTKl3KtKnTp1CjSp1KtarVq1izat3KtavXr2DDih1LtqzZs2jT ql3Ltq3bt3Djyp1Lt67du3jz6t3Lt6/fv4ADCx5MuLDhw4gTK17MuLHjx5AjS55MubLly5gza97M ubPnz6BDix5NurTpgwBS/wOwOrXrga5ft459uqtqgbJx5459WzfthLlh82btm7jB4LOFG/ds/HZz gs9xQye5fPZu3tCxF+/9mnj3gtqV/jd3Pp71cPPhhyP8Dh68evVGl8ufLh02/dXA7dP3Xr+/fvz/ +TfUd/BZ1xtR8wF4HGrtiVSdggom6F90AI4X4YMUSihhhfXxN+F9/x0ooIceQjhghguaSKFCK4KI YX66ZRefcP1piJSIJXYIookdmdegftWt2CKEPgZYo4Ac0jjij0wGySR+D6q4VJFD5hghkvdFSWSK DO6oZU8JavllUdwh99uSIRVppJVCAinjm10mud90bDbo5JoLqekllkHxVyWa1rFoZIAvxrkngncm 2hSJdh76UZtzZnlmoO5VetxuKYYpZXY4Tjohd4MGidyA0v25qZxxirkjj1jeSWZ5/qKKSKqbhuIJ UnibulonqzwWumqOfj65a5SuGsiljkdZiOSuvP6oaqi12possrTGF52N1TaLka6NZknosXr6Cu2W MTrq7LFXmqttny56++2gR9a6YbanCqVptGQmZyCs5Hna47khtvuhfaCC+uTA5AKK8JAJF1vvmDth yim/8uF6qZidymqsvhyfmC2kg12H3pkFcqwafCOXKuuolOorsW/ilansxMrVDPN2ENdGU8r7xpyz zkAHndZ5Ggtt9NEXA1d0XksjbRfRP98YdWtVTb2t1Rk17bTUX7KclNbrYW1vSWD3KPbWP5Vd7pQX nZ32SGpvFDfafZ49t05uH8z2/q15U3T3VAUGfjLOhMMsG8oWF16cz6hh1x2u7J1nUd/rgik3VJSH ndLfVbsrp8HpxtvwuwCPDu+17yrLqN9Zz8hR5jfBDqdJnFuFrcCs3j6vwstmumqvqasrkeyVTwz1 8cgnL7lHxM9EfO0VQW+7uxX7Tjqk8xZq6u6fX/8s668/1PzkaS6a9fgNof8U5KMeOKnFerZc7+K9 b3xxhv6uHP2j6TtfPlM/U54AB0jAAdoGTd8DVu8UJa1qaY9P3aMWutymPghacHNwA2D/XLcVkF0v Ve1xGO+A1y1e+QpFDYRIBbW1wom0sHjm46BWPIiegDkQXeTiHrE89z0bXiiF/g55YQ9h8sILOqWI GexgmXoGJZkdbolr0k7JbOYemSmJism5ThNjNbwk6s0lSETizqZFN6CE8YtgJNuJuHhFI+ZJiu/J 2P7K+DbqeGll3vnbFH/HqUvdTFBmJF2rtkUrEomKfHT0yRkBpbsuwouFH7ygGPm3NhMScj0hJBgi ExmxN2IojzmLX+lCJz69RY1h65ok8xqlv7bVT5Ol2iQnczImkKGyjV2qISmDaEpeNhGHaOzkgpwT zCC2L5OxnOMsaelL08HomcS85TMHyRDV8fGawkTm/Fz4Sh1ZKSKqXKYjq6ky0WkOmrCEofCKh0Ij hlMjqmpR5rrGSjdOkyuf/lRT5E6msX2CElyJy93g9jWyJ/rrX+Qslz7zZMmASROTvwslDzcYyFwB cZwJ86Y9IeoVXZUIdDtcmw6bhENrJqmd6uRmQq/00G+uDkqPBCSfIPbSl3JUkb8JzjEnx77ydAh6 76xJuOhUOnGR8kUbclL2hGWoFYLNRi0lFMmsCM6AWjA9/dJaUFuy1ct1VEXcMycEdzdUPIXVVkm9 VSqptVCZppR2asRKV+H51WV904cxnSL+zIRXeen0oK1j4VR9ylBsrmSRVymgYhe72I4CVpR9bVb2 SGbXl90UVXcN7EZlGdPDxjWxZOxgCWE6UaYizKKnRdX8QvpWim5WmcXE/qCDZhha2u4HdQsDoSAZ mNHQGfVerS2sYel62TTONitz1SxyrcizP2axn+7r1x8FZzKq9rFwWlTucF3p1pc49VVchF/6HCfH 6vrIoBgVJ1fvuUrhevd/Y8vtCDl6V90pML3qVQlN4csixvoXeZQMbSu3S0KSDuys3c0vSiSqVnaF 72umJfBHmzq73jZTwQu+sFfjq918gQuYjQMd7q64R9di+LMmviSCOuzhUXY2Urnj433BeeKTiO15 N+IuhHUb225GUoJ2qzGKVchijw0vuQ8uFmvde2BJfUylQuYvfmkMwKlJzyZwDGsAo/s+nbpMulSO coCjZzck861rVz5i/m3FvOGric/M1CEaWuAMWzYXmadopnNl9AxlO6s4bIBV2vL87LP/Gvq/hP4z Oqec6NdiudF1Zi+RIY2vilJ60hu88aUrXcdNp7i/d+NzYswEZgkPTnKIE5nhAk3gTWMts8WdpS0v yqWB1tSHjWSypzktaV0reGbydO1Sa/1jBu861iutnKgdwy2jKk1e8nWmsY/d4/s9lc2sXXI1HdZl 90mqbMsGDWKjHFJRahqtEfXcp3c9biHHk9YGnu+SH1ptQre7xsOCN5yGLcEfr/vS9z5xlq+73eaa F733w6KvgUYgq2q4wdROsIOPZtLgOrrPEUd2p40G3EiPOeMaV+TW/nB0cYuHGeSCPrTK/Us3J2Ic 4ijv9cZH3u9/PzjmErc0x2F8cinjnNeWQ5rBENxqzkKa1P8s+KrPx+pY+9NrlOFZiYFOXFcXe7gf hew4z+1ksaKNWOWFetGxTWJB8li1L8e0vzl5wtLKHOAlRPAh6w3RIPtxtJyBWimhiVIQ046w9CNo FUVWsX9SderTG6bZTUn0bXMOzTTL+9LEftW4vx2uu0VW2zN6O84fMFNeljm9C2v3RYd7LHFrGj0V P9OYgPTDl918jTxochmOGJJrZ/TDO4+Zxw8x3fJ7sY0jePtIyZ6lqT39/5TsdyL1HaOvvvtRLUM5 bT9ypDIJVjfz/uzSSJZV+IvKKRSDOZ/Q+83h7G0u4hsDO2erFvvZT2aBJVvzWZN2vnNeOPiPexoc /0r8MzN2diRSsAdto4Q/65R/+odkyucW49N4ezdGmXdN4pJWpwWBqPdpZtaAOLFyHhhgsiNqlhV8 LNM+JmhdFBMXHriCiiUXdPZdtbd/P6d/IgcXegaD8zSDd1aDDiiB5VN9OriDQdcWfCZEtceBomGC znVV5dUxb9Z0TndqWUZ5OeaDDpI3SBga9rdNeMV7zVR6ITJhXqdBQkU2FBSE9KVu7NRGo9d8KdZx Mmh7d/R9bZY010V5j4eGdUc1uxRARaVSZyh9QFZyQzhBuGQ1/ntUS1cHagUkFdbVMiWzfmVhTcHm a2JYVbUDeYEXfHsTLWSFae6XKs9Hd4l3gRMYg3LlUPpGTZyYaeADI/eVhT0nY4r4cBJmeTnnWJHF LLLoP6qIf9dHinHoiTw3jDrXfEVjZawocW11eaJ1ifSCW2xxa5W4hwkobPg1dDzXi/1Fg8voSdaH L6C0X5Mofw/0bdzIElgFeHkFgBWmQuiXfl8GgFR4jMAEfwkVjl/UeOk4gM0YjfTyF/0ogKUojgyj jJfHLGtYjgSId23INGtGFqsXQskYgVwYM9N1SmZhgebEkQIZkRLpjG5oPKH4KI0VFSM4WH8FhXfB gi4pQGsx/pE2l3bMA4Z6aHQTNzQiiYojeT6veJMeZ49nIYJmGJRAaYtCOZSPNlt9M5CaoYTfWEVS uYS6t3ANd1BOiZTrxZQ6NoM0pI/TF1lqN5azN4ZZqZXquHyKVpAf6WNuFIAPyXqzmEOCcpYzKVuU 1H62FRh+go/3NIq8ZJOHOEh2aUwvyVggWHVs2RcVB5anM4YUxXWNIx7ENmoINSJhN2Cdw5e4RnXZ Vo8hpkf5WFlpNhf+52/PV5ho2ZJuZ1hDFIjDOG+tR5B4oZf7WHZRiTmAcWu5CZC0eZGSpo19pZoc 9pPBCGO/tz4COVgKJ1nuiEvwyI7JRjBZhYKlaZpl5nQM/nJLxLmTiHGW3SkWqVeLvdSKxpiTkgGe TgN1VBiK/OiI1AeSQqN3GqaQGqmbe3aY+jloiVZ4pTY34XmeISOfYlZEAcqTjEmg5DZkOxZ1Crqg A9iJfAGVJSho0geag5eJIbZFhFedB7qX/Nege7GF7nmKYCiYjdmaH4pcDNpibdlQWCctNimZTSaO wngZATcru1lOB8NgurRL9TllJIdawKiFLVqcCapQwMd3bIVxNLpvSMdvtZGjSQmRHQl0KBSXq4il /TZjhIijRzpzI9qa2IQtWlqkxQWHj/mbvRemhTimEdR91iiNJ/pywnmlFqYzVMoTK8pVzGk/1/en VDle/ixJjAbVoeN4nXVBn13pc4TKof10mZOhnkI3eVe2p6Jolm0DYOd3XgHlnxiqZnI4pVbWlG4K e4rYlJy6ddAYPAiqoysWNL5nlNOZi5NZjPLXqfx5crB2fwJaZfsZrIoKGUBonBZZq29EmS6Gibta VXiXq1+qFJQ6pUJoq5iEiI7kUoUKaM16ZMqqhn26mnyqp3VIlgwVfe6VWSIUnZK4d9rKV18xraXx gLS6rN54p32IZ/GYrrO5eIjySVQpr28hrI3YXjRpbVUpdevYnE+4r904iApTmCR6o1bYg/EHc3MZ Yceafd1aSodqnUknrcf5qxc7sBV7cweLpiHnp9vq/oI8Oi4uOo0nK6m6p2oN67Bx1rI2qCQBGK1E ZLIzi7Ip26kWyrKhuj6IKVzV6LPvRYRlGKHGSmZvZrSHpbMmWbD8yjsCq4BBS7NDO14uFFRWW5Pe ZUCumK9s6noye1ODhj6YKq7Wiqxwc7TwOLXqCJOEeng7FbMb+ZcQq5hehJN3Obhmoz55WLWrCrWw 2rcal5xRi7H1SnUbO7dkWzfHk2GjKp5vJ5OC+3GdK7nmmib0akbJo18P+nkhx7mRW66fm7Yr+zrN I4srObZMR7D6uZHk2LiAG6LVCoPoGTsliLOe+7thkbucFoLD57UUS7iJS5KPdbkkm7Pg2LR8q7mb /uud3ii0uxu9cuu62ktEGPo8eKSxixuSjVuR1cq6RWaEcUuGhmlcFjak1ZuBr9uTcLu9u0ul4TqS tAtfSLW8Jcu19su03tu9gJuj+wti/euPfAitW8u4oOu2p6pcAZfATTWsEEdM73e6xbuUgauWmMe9 4IXB/PccS1u+AhzAw6tWNDW+bxlRLvyqd1ucJgw5HGy9vuiooitTSeWHrNjDYSu8ARuy2SuBrbqu VaqUXbuWirtoSqdsMDtpvVqZouPDYCKoW5pNFqvC3/vBsNjAa+jDP3qEIolKVgw4NzyJtou3ytvE mQqkt6q0rZqxCznGMGTBD2ykXezGVDx64NaZ/kZpbBoMxu0roiichOo7wbl2rlkLx5PrTsNJuFU4 vy44V7aZloB0iWL8xb7VucZ7wvX7NWt8u0yTXMUKvwibkZ8cmoBHt3mVyWJ1xkcpVAyYnT+ruGKL vqxsM4VqwQVKwmFrt2rbxEWIubNMy8BMtIDmwZBLwPh7zInlynPbscOswzLMxNBsO9LMkCssws+c zY4oxCm8x95sL1LYUymoOCTmcozTztgmzuOcvtc8hLaGp9JGYaaYe/gGz/HcuvF6e+cImbh5fI7s btSsgonMrR0LvdbWriZpgNsYkMkkp6cyxeolZ3pxydhclWDEbaA3u3tI0R6VzKaB0RndqCj9/rgB LEI2hXYaVWDcMs+ketCVHMzkHLrJe4r1Q3u1JtL2atAmPaKl2s10zKAiVoFd+Cvq5pEYSEcMzZiW 6rspbcCTa3An2MoFhUdCMrtcts2VGtTLSdPyTNXYK4jg7DzNy3Df7EnOetZvKnRr/ajJ6tY8SHNj bcC+TG15jU9xTdei1XJ97dcsCth3LdiGTdQ4fbd5hEV6+5y0wZ+CEyvr+Nil1nKjfNmYndmYzcyf uIjxwosVNsir46WHPaEouW8XKmO9+W4TDUQWXdp0saLlt0A1+o6zqVSwBMpaB9u1edpu8kD+WWkY Q3ImvIm0zdt+IdtzAty4qssO2dr0N8BR/gdmzNWEmLmwSfc4KhnVWxSwGWp9T6fdHytGyh2GPa3a MSbc0K26Ml0YTH2Iv6V5uFnFxfdDeYrUEabBvJmnBiuqqeV5T3abNkQlYwW68SlVv+irCp58+rxa o/W/AI22F9XU/vyv0zWVF+LYNkuSvKzOCufV7NcuE1bcRFWABaiQF/mJzFfiaxXR5VzIyF2QmkLi vmoqYzXZ5Dvc8Op8Oz7Fkm08IRzjYGGBbUXgGBlvJ661An7d3tYrOcVksZjTQv7P3vPbsYzkYQmj Sg7ggLwkHr2k5D3lHWyi0O3gS33m/93JZodSvBXfNzTBYl41Kwkk+s1PWM1GNyPeTsTd/nn+n3ZO IzZ7gu2cXbwb54ZOvYee6FsJnxyuzpGtaoan6Juxv/jN5WpuIT0r6XuGxl8e01432y+u6Ul62p1+ gB8zd6KeGX1Ke7zYbQBr26keGav+Ps/dpf0a6vN5zvnT6IKXRZYyjwiX1ldJ2X8Ksu3txRKKtp7e 2bm610BLNS39MPKd6RyJPbYeI6ON5lXcj+Wd1PnseQVzponkkSzeW+TewFrmlt8ebeiusrGDOdW9 3YUH6FLkMk5oZ5G6nREO66Slpvkex+bSJtF+xU8d60hbnhmumdDJIcF9j3MdRebH8J/CgaVr8Gh8 OpSV3sRY43YlucO24kaO62cm1hYP/qvM3qRIfkLpPn8j5ukNg+K0xKglPyU83eYAFcVEnoBHbeqo iVNgPfM5NudURF0XLuizp+GVTZ3znpHzKFUkfbUODfRSX2jOPvXL9PNWn/Xf/fRaz9tRrxiPzmUV MnDPlahmH8eEx+tf3/XufOD0XZZvb8887UxcLu5av/aWuYv22nmymdow/axsz65PST2vXPcuTpj1 h27HLuoLTBjo3I5Nnk6baFXaR9DjGPhHxvWsSaSrt1S85dIvbfk1h/mON/j/7fJmhfAJ5NO4SvrG NOkTSENuCVw9JPve7vrb9pRLFIldnfSS7ecY3vCKg/eYr/m4f/zIP2ebbc6XnfyL/oqXSYzYhqHZ LLhzxhz9/R3imTuv0C+myA72acwc3V/XRYva3fv4vOxTqDaF3C38vN/4X6vFSpT+SI/17BF477/j Qe79PF9J1wsQ//4BEEhQ4MCDCA8aLNhQoUGGERMynLhQ4kKMGR0qTNjR40eQITl6BFCS5EWRKQeW ZElRpUuVMWXOpFmzokaOKB86hBlR58+bI4XatAmT6FGkQ0Xq3LjUacOeOE02FWqUYtSbWF0yTUrU aM6gK7vuVBry61i0aWNOHfkVIkmQW6mS1Xj2rNqOd/Hu1ZsX6tW1TxFibUuX6lupcD9q9ctzr8y+ XPvOnKz48WXMg5sinmg17ufG/nIdN85sufTjyjk5T77bci5QzWFXsp1NG6dsw6ptn769+C9upKmB 8yZO8+1Vlr5ND2cc1DXp4sKLB4Zct7dy0L1hExSdEXBZt4q7c55etu5qtdLnlmdvFmrWk9ih4yYP dqj6rvjbL3dvfT1/2Qj7TqwAsfMMvtBI0y8tvVDiKjiv9pOwPwL/qw+8BDP0Tz7iFpzQPAQdpC67 xEo8DDr0DAwLts3Ya825647Cz8MP8RqwMBxjtK9FHnsE8TIaJYzsudksisys3Ygscsm22GppquQK sirJKGuj7ckgi+ovxR+Ni7DGGqv06bkoydzNyJOcfLIzJbvkC8wvp8sSTjoh/lyMSDHPLEpPwer0 88/qypvTRUIBNfS0QYU8dFFG/8ssUTkLbXTSsSBtz1JKM5VRUk3dtLFTUL0MdbhRS0UNS1RTVXVV Phll9VVYUTV11jRjtfVWXFOldVc/MeWV1F+DTQ8zX4U1FrVjkU12WbR8LZZZaOOMNqlnpx3VWWuz VVTbTbn1lsNhvxX30XFFTSk57tYEsNxeiWX3XQbhfWleKTc6UN4/scV335qq5fVI75zyl1/K3CX4 4HMRni8u0TwbWOE+P4V4Yh33TU3NJCummFwgN6Z4SnXHlQ6xntr0mDd9T0b4wAu5BXi05R5WOWWV CX4R3iED3tDRmiVWtmd+v29+98gyQ85TZo9pBlpeodlFeumf34Qa36bLfXrqcKPG2uo+i04XolZD vXrrSg0mm2sKnT67U6XXdhnJsd2Wu9m5v42b7Vzz1ntvvvv2++9Yza4727sHNzzrw6MtPHHGu22c 2cUfl3zEyY2NvHLM183c1Ms337xzz+kOffT9QCf9dNRTTxxw1lt3/XXYY5d9dtr9Vv123HPXfXfe e/f9d+CDF3544os3/njkk1d+eeabd/556KOXfnrqq7d++IAAADs= ------=_NextPart_000_0000_01C71122.93BDACB0 content-type: image/gif content-transfer-encoding: base64 content-location: file:///C:/Documents%20and%20Settings/Klaus%20Holtz/My%20Documents/My%20Web%20Sites/mysite3/THEORY/info9.gif R0lGODdhKQIJAfcAAAAAAIAAAACAAICAAAAAgIAAgACAgMDAwMDcwKbK8IAAAACAAICAAAAAgIAA gACAgICAgICA/4AgYP//wKDg4GAAgP+AgACAwMDA/wDP/2n//+D/4N2cs7OP7ipv+T+4zUiENpWM QY5eQqBiemJPrB0vvihmdgBFAEU+AWooE4U5akoyhQQEBAgICAwMDBERERYWFhwcHCIiIikpKTAw MF9fX1VVVU1NTUJCQjk5OQAHAA0AALeZgYSZtL29kH9/YGBgfwAOABsAACgAAAgJKwAdADkAAAAA mwAlAEkAABEROwAvAF0AABcXRQA6AEkRERwcUwAW/ysA/yEhbFkUFABRAEcaahkyZwBhAAAx/2EA /1MgexZDZy4u4iZZFlFGBGguSQdSj2oYuJAjFQBT/6MA/2pKEnUzbEpBmjdlC6QsFYMfsU4s/yBR tghkkm9WC1lDrTZyErAzFwChAHdfH4lHcbBDHLctfQCGlXpuIyafAHOpAQAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAADKAKxbASAdwpRScCSqTAqUiTZue0R1kP8AqABx/98A/1aRSjRI+MwyguRBcGjK ATa8QgCa/5Yit4V9MyW3jDZa7Vz/AP9IACKbokLPTcJYUiDTlaUk4HNWtampANBvPGefWInPC/+s AKcu/uJZf0zcZ/8Y/zp9/7HQGMf/AP/iAN+aPVaBn8ZDuq9xiziiydFTzv+aZUbK2/9N/8jpakze 4P+Y/9/AgunspfX2zf/Q/7GsWmORriJMZY1OP1BwcND////n/2lpaXd3d4aGhpaWlp2dnaSkpLKy ssvLy9fX193d3ePj4+rq6vHx8fj4+LLBZoC/eMbw8LKk//+z/9GOo8PcN6CeVHaucHiewYNkv6SD 09E/Mv99AER4IyRfYA4OLL4AAP8fADE5ANmFPgJ3hbDYgVYhHQAAMIjIs6B5AMDAwOpwgVHxaf// gJF0zf98/6L////78KCgpICAgP8AAAD/AP//AAAA//8A/wD//////ywAAAAAKQIJAQAI/gD/CRxI sKDBgwgTKlzIsKHDhxAjSpxIsaLFixgzatzIsaPHjyBDihxJsqTJkyhTqlzJsqXLlzBjypxJs6bN mzhz6tzJs6fPn0CDCh1KtKjRo0iTKl3KtKnTp1CjSp1KtarVq1izat3KtavXr2DDih1LtqzZs2jT ql3Ltq3bt3Djyp1Lt67du3jz6t3Lt6/fv4ADCx5MuLDhw4gTK17MuLHjx5AjS55MubLly5gza97M ubPnz6BDix5NurTp06hTq17NurXr17Bjy55Nu7bt27hz697Nu7fvvgCC/wMwPLhx4gWRCzwufDlz 5c6PF5dOkHr05M2jG5+uPTt34s+h/v8ez1K5cOToDYqHnn74cvUD27+fT784dvjz2Venfz4/efX9 bTedgNxVd5553m3HXn+t6RcfQuDt96CEB4nn3oQWmnehexniV598G4b434T2heigiPxRuGF2CL7m HYnJmagiihjCCGKKGna4oootejiijR9SaOGKGta3I4wNrlchicz5iB2L1unI4YMZErgggfEp+eN9 PAq55HvyDZmlbScaaKWRaNboH5JHRohimBKKKeOWANpoHY1Fumniiy+qVqaUa76ZkJRFejmloDEy CSSbdAY6J541FpomjaexOKOIcCb6paFBRoomoIVayuh/Dv7JaKiSHjmpaeAl6CqC/lHCWuGrY2pn Zn60ZokelBzmSqeopnYqabAlUtrosVe1uuuyVjb5nai66oclstRWa+212Gar7bbcUjVtt+DadWe4 5Mo1rrHlSjQtlnzKGq2ZzcaqZ6LfhvfurFfeKmdo5zqXrkbDaiowsY/mWSu9SLYHIqFslsnvtwIP yq6z/l5noK7wwgtlv6L1uCq0DMfpY8iUOmzsrlyOihnHTkLYpZEmPyrziamSpmWrS2bK6ZoLuqzo pogCDSi6XdlbHsQSN0RynWmC7OHNrA2JM6J6Ln3huX3S7POnWzd86Vd9FjsSxRDtG3Gb8N0IZoqa Qk20Z1Kr6inXMDs6o9ZCt0z3/sthhX3xRyw7ZPbZsOo479xB34w0aGJW/XTNqartMN4j6x33wGIN rrK6i5et+aUGI3y5420j/PZmN+q8aKesSxv5znd/vbCTn09V+6pKd67uRFYLeXmxVqs9mrMTVxl5 8fEKqCztfq8LcZPFp317U9Pjjm/1SldEMOaVr17w1/8yhD1JNSfLe+5+Z5S+4NEDCDm0tka7se7h D1pT+VaNXzL9GOlfv1b+60gAlyInBQmNfwD7nwJRUsCDPWmAEVnfAie4kcadzXoggSAFN6gQC2Yv JRrkoAj1droSAiyEI5SYm5SXNY39Tl/dORB1nqO9B87vVcnb2NUYRBQPiu8k/ihM4dbg90KZ7a05 CmObER8CKu9RLkxVCyIIu1Yyk0hRiBELWIy2Ny8tJSxvu/Oi3bQYtENh0IrhSSPZSkdFwF0Ri+Bb YhdjF7PYzcyGFhTj1djItp79bSZqVKPKFvfGOMKxgipcF5PGFaqFrAd52CPjGOWXsZRVTCaBpOHm HDm2Q5LPfgmL0qdE2UHT2S2G15tc3SZZxQseLZNFFJxIJOhJ9fHRiKpz2xdlt0mgyW2PnDKZtAwJ Qli28YOAq6UVnWhHSapqaKXqJe1WRyxTWQp+L8kkJ8sWkkIq0323KlAOWxQvjD1JfstyoCN1qMkB +UqGyhNngbIJz8AR854U/qHlNyd4J6Oxb50Z3OchkbZGJkqtkN4U6IjsaU/xtdMj+lRouBB4RkBK lJ8U7eFF69fQpER0o9TKKFISClLZdJR6JeWWSAlI0pT6aaVMaalLBfjQotXUIjCFiT/z6dGbcjBs H3UKUPV3UpsMlXcyneVBe9pP4s3QqeZ8qCbZ2U4DFkWCQe0pQA2aVDduVZZHwWpXA0rNnzWSmT+b UxKJdFUmmg+siRypW5HZVrhqFHZ9nJX3fnlJPdaRJ9Mba3kiiC8CEraDgoXoYe9qVkPNbnaO3U/q mMZYbkolkvEz7O6OKRTMVpZIOLyXf1SZ2T3qUZ0/8axOiopPcK52pap1/mVnz/fZ0NEtmqQF5mhl a0KL8vS1PvUlMRMbwZ2C0oXHvWRYazgUL8bydwm65SKz2FqaGI+QO9FmKTkrTXoKMrl7Da9RDppO 3q6Wb7pdFMGCldu+BsWHJwOudmV7OBICEpZDE+8px3vcBr6XbFPFIRJBBmCnDiu6OXUJfP2LE/xS rIFmIy5XHSy8JZpTs1/al4RXs+BEUvjDIA6xIH2oYRGb+MQmhlDgNtxN+0UYxTAOJL9U3DkWr9PB rixodV+JX/BVuLvvVeE2ydVh7sbEmMc8LZB5nEbO4q+iQfaZhtNVZOFat8lDtqOVr2xcEj65t6nt L3gxybiuMfgmwRVu/jm3fN+Tkhe1UGbyh5ML3yPb2Fzuw+5rf0vbPW+2zzkJ8akGzKcr24y5TyFq VBQNWArv2LxADCwk/TjPvoYWnZi2HaATvWmhdrrBPbbvdun55++hF62se6a3Ph1TpN6ZrIvNLpJF PeZidvrJ2HTvEU2Zv1gverEJPq9lezJfI7O5JFkFnZok2zTdBm+/UKndq5E91we22q6BricCpd2S ZBNOtIgTGeR2ydqRvjgrYnXotC+S7mwLOKjtVom3W2bbcCPu2enN31K5clTPrVt7+0Zz88oN53n3 L4S9m+7jvLbKf8s7zVjpsr/NrWOd6o7geJSwFLd36lSvV7/hwzhb/jLqcIBnsH3nHOXx3p3pFAY7 cy8fb8ln6saZu9vmaKQ5AA2ub5xHWudeEXmiY64UngOdekaXuc9tfXSzEJ3YT2dp09cSdVcL3bX/ rbjJp+6WqudOr/n0+s8py26u9xDL5OPfvOONPpl/FalmvyqOlVpA2Fabxkv/Ibb/GXelO7qbWdM6 rS2XdN/e3aF9L/rfx3ZDfXr26mhmtSUTT/FiI9vygy+srLMaW7BTPqazZiDm4wxpO0Oc9D7O++et LvG0h37QF15yt2Wc5IpXafWX/W4x5/4xHzf69Wcmsepxj1PdR3rxdR/V0gV9ZkgVnvhQNzAame/i SSkrxtjPfvMD/g79nt++waCsL+ofzvvtH7v70R545KXcSvmiPY5KHj76EYnY+/W3zgJvfe95OX+n Tzv4xmZ6akdnAdh//PZqb9ZsjVZqLVd6BlhzLYRcqAQ8hMZQNsZonMZnDzhFq8Q3Z5VXp/N8iDZs 0SZ5G9hJQSJG5ZNr1SOCDLh3KPWCJ0htKcg96bVCmad5HHh4JUiCM3h8lONOlbRbDpg0O6h3VcFt PyhvqSQyTkiEj6ZiTGeEyXJuS8iE1KVEkKJc4yeFoud5OweGV8hA06SF0cRwFTRW+hdxpzeGgJcv 19Es7lRemOKCXOiG/WeHeAhS8OYuB4M8DdiGe/gjHxUzH+dx/rI3iOPheCv4gcumhYqoUhGWb2vD V5MFbZFYLukxOpbYgZCXiY2yiRnWiZjyiKBYF03VVLFHgVfSio03SwX3h/jhXKeoF9Vkg+h1iHzF EYaYi6gGZrWYOU9DhQE3bg13ee+zcoXDcsGIZ2zUIcKDg7vmhPLXjIRBKPVSiaKDhoqjh9ZYGiFj OFtIOrCzVt+4iMM4RNBljKUIjOc4GxwnXrpoal34jlGDcuH0JtiUiuNkj/74jwAZkAI5kARZkAZ5 kAiZkAq5kAzZkA75kBAZkRI5kRSJe9p3kRiZkRq5kRzZkR75kSiWSiA5khwJFtWoYG9xkt12fmSh kgKUkjDZ/nUsORYuyYsx2RY1iYVFaJKAF3vyMmk2yXji54raNljsw48VyCwDFlUF9lQYs4a1VnwR wkLyck469JQtloPoE4EtF2BMKXhll5WIOJbuGJX0JzbqFSjEdTu3CIJyo4taZI4+GJYl0pa3RDMI NZPVdlpPVFaUeJZeBYn59mU4BYvKBk065YMraIq65omep2Q8WJiPGIQK90t5qZXINEwe+JiYqYGK pTqs9JeSCWvVl4Uap5iTqSaamVuP1YEmCHe+qICHQphBuZNIqJl+uZqd+ZphSY6mWDgvKZaV9jpD qJN0BVqtCG6iuEsDIm62+Zxf9yH9tCnLuSdvFEtSGYRN/olbuymD9OebofljnhmcoaRIT2icSOiW VOOczAlZh+M/l2mXkUWcydSdtQZgx2hvZTmXtcmCjlmfpIlLWraSh7eYF+Qx7biFaAmftXmDePVM 9KlY9tlGtJhqlZmI3plAqDaPdEma7aWCK8GWsdlEd4RW7ZWhsOmgaeVssQmgUThq7WmiyoaCNVee y0iUDXpy70RoDUiGRwmUwyl9/LGdPCqdGGhLVDmBcxiH+yh2/Kc+O7o8/hJgK0ejRXOTI4elGMpv PCmTXoqTetk3XQqmX5qlEwpAY2qmZLqmL8qlYKOlapGTR7ilWyGntlSmVAen+4mmb4qncaqndgpE NNiV/kR6o1j5HfhkVUk6MDr2ZkpZlEs6gtmZnEz5lDREpU66pxpTnCxkqdBjYJmKnZwDiJXKpJ/6 VJ9oljlKjyWKNlPqQIPjMb3YcdjIM0r0V+lJl5A1lmvliFeEcKkpo8Dkqy5Kp8/YosOqliDXoQHl V/8pmLhpYVomTK2qjQnqmsuaZbqaOH/krN1Dnv1DJenYmNhqoRoKndTJrbI4oNkqqcnkrRWqnu1y fgC4Rac2RzCqovUIrNZqhoPJa7jqrohWXwFzicFaj/mKsF6yNNJIbh13rrBGWs7Fj2S3oNoqgWm1 Ro7oS0g5mgeHRONqWlkESQGandMIhTBDssWqsPfC/rD9+kgCJpwQKLHHep4zaj31Skmn9GDa6p7M 6rHqirLteIYrK6kua0b5iYnheqbs+pYvC7Asm7A5SpzxSq3imq3xKqBOa7PeKl2+BrRHa7FPep2A GbRs1bS7eKfoWoY7g6DXuqLgKqErKkl0Czx220t167AWurHXCppf667z2LUsWrTjGbhZqKxFq0Ek Spau6bdly4tU9ZM7GqlSKqXS1D6LGqn2sWaVZjGUyyu8eZR5Fk5ncmFUCmc/G3bXRbpE9IrzpIei Wlyr+5V2Yqiv+6tpmqd+mhax63S5+6e7ixa9WxaBmrrCq6fHy7QRZ5Ik2bzO+7zQG73S+7wiOb3W /st7FZm92ru93Nu93vu94Bu+uWFVpeU8PgWzVFkv8NQdMVS53qiDnpuyD8YgmWtUfticWBe/1lmV 5ZWpE4dHz1Kkmis79CvAeyGrPPJI0No9BlqDtiqtJzd5Q6uWzmRUyKpOXJmsZwvBcyos93qzlUmb c6FLipIjkPieNvutSNu4ZGjCzKbCW5ttcCsoLny1H2yuRilme6VKrBm1Z4F/NdywSgKiounCkoOL n1TD/YrDTGxRfguNNrzEfxWwgkqAlbi41gphgQHENrxC5rm/wMl+cim//pszUcyOJwOWiWlG+eU4 lDnFarxMFHrFeVa1QgYYXEyEFYafalWa4Nm3/j7asEKIxA9rXdrYxnSssyBHxXJsOSKrwjzLfn9B wnk1xs9KmziIxhwsoUJ8uAnKyChJrrc1uL8Yw9m0ZZkMRhMcTHj8i8sZjXxUwYzLwlXcyeyat4F2 wXrMnJ9cyMaJwvr1xnPbpiMXeLnyqCkXwOiUPngJh+L0vq6VTpxrvqxrc6ULblFlqrSbj898gaDK LgRclZ6LuWUsvuZ8zuiczuq8zuzczu78zv/YqaZ6lQpyvgVMh2QMzRNGqRYTudKBqfoMuaTaz7Ji LwAtf9vJup76z99cvPzll1tryQiMhu36SY5yiL06zMR2sCWa0f4aZhc9og9syotBaR94ojH6/kVE HMor/JYRqqCaWsUtfWC8DNMIHcVm+NIv5NCVZcskjX87DMinLMVAYrAhbaw+OtNFja27GtNJTbBL /bZNzdPNxZ74u7fcJVW1G9A/erKPbKOb6hMzFLLVScZhnVrxdJdPy6mpehfcKZtvO0gJDM5OrbZe XdarrLQWfLNvrbd6vX7caNVJK5qJgZengra9DLd8G6Lf1jqCO8O5zNe+48mQLWyBfaFxnbaIIZ9B PUSbycprjKwnSrTDh8VcxD0ibMi4eNovk9pbLMD4jM2EdMwy5Km1kljqS9sHZrtyuHmju83WmWm9 /Xu/TdC7LdygC8/KvdzM3dzO/dzQHd3SOD3d1F3d1n3d2J3d2r3d3N3d3v3d4B3e4j3e5F3e5n3e 6J3e6r3e7N3e7v3e8B3f8j3f9F3fzR0QADs= ------=_NextPart_000_0000_01C71122.93BDACB0 content-type: image/gif content-transfer-encoding: base64 content-location: file:///C:/Documents%20and%20Settings/Klaus%20Holtz/My%20Documents/My%20Web%20Sites/mysite3/THEORY/info10.gif R0lGODdhEQISAfcAAAAAAIAAAACAAICAAAAAgIAAgACAgMDAwMDcwKbK8IAAAACAAICAAAAAgIAA gACAgICAgICA/4AgYP//wKDg4GAAgP+AgACAwMDA/wDP/2n//+D/4N2cs7OP7ipv+T+4zUiENpWM QY5eQqBiemJPrB0vvihmdgBFAEU+AWooE4U5akoyhQQEBAgICAwMDBERERYWFhwcHCIiIikpKTAw MF9fX1VVVU1NTUJCQjk5OQAHAA0AALeZgYSZtL29kH9/YGBgfwAOABsAACgAAAgJKwAdADkAAAAA mwAlAEkAABEROwAvAF0AABcXRQA6AEkRERwcUwAW/ysA/yEhbFkUFABRAEcaahkyZwBhAAAx/2EA /1MgexZDZy4u4iZZFlFGBGguSQdSj2oYuJAjFQBT/6MA/2pKEnUzbEpBmjdlC6QsFYMfsU4s/yBR tghkkm9WC1lDrTZyErAzFwChAHdfH4lHcbBDHLctfQCGlXpuIyafAHOpAQAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAADKAKxbASAdwpRScCSqTAqUiTZue0R1kP8AqABx/98A/1aRSjRI+MwyguRBcGjK ATa8QgCa/5Yit4V9MyW3jDZa7Vz/AP9IACKbokLPTcJYUiDTlaUk4HNWtampANBvPGefWInPC/+s AKcu/uJZf0zcZ/8Y/zp9/7HQGMf/AP/iAN+aPVaBn8ZDuq9xiziiydFTzv+aZUbK2/9N/8jpakze 4P+Y/9/AgunspfX2zf/Q/7GsWmORriJMZY1OP1BwcND////n/2lpaXd3d4aGhpaWlp2dnaSkpLKy ssvLy9fX193d3ePj4+rq6vHx8fj4+LLBZoC/eMbw8LKk//+z/9GOo8PcN6CeVHaucHiewYNkv6SD 09E/Mv99AER4IyRfYA4OLL4AAP8fADE5ANmFPgJ3hbDYgVYhHQAAMIjIs6B5AMDAwOpwgVHxaf// gJF0zf98/6L////78KCgpICAgP8AAAD/AP//AAAA//8A/wD//////ywAAAAAEQISAQAI/gD/CRxI sKDBgwgTKlzIsKHDhxAjSpxIsaLFixgzatzIsaPHjyBDihxJsqTJkyhTqlzJsqXLlzBjypxJs6bN mzhz6tzJs6fPn0CDCh1KtKjRo0iTKl3KtKnTp1CjSp1KtarVq1izat3KtavXr2DDih1LtqzZs2jT ql3Ltq3bt3Djyp1Lt67du3jz6t3Lt6/fv4ADCx5MuLDhw4gTK17MuLHjx5AjS55MubLly5gza97M mSaAz6BDix5NurTp06hTq17NurXr16s7y+4IYDbG2rZzW8SteyLv3nBDoxQe8Tfwh8YTEje5/DjL 0sNJO0z+T/rog8Y/DwQtkDtu7gS9/mu3fv16d9HKt6M/39w7e/Dia4Ovvt48Qurq7ZeE7pz56eim LUTdd90VuJ1Byf1G4HnqFaSdgQvKB6GB1YVnIYIOXlghhbw9uOGHHR6oYYgiYofdfyeh1h9IsaWo 2n0KZadhgdmRqOCBIZJY4owfZrjjjzdSCOSPHA6JIZE7tuhfaive5hqArPHoY5IZyrggfRjmaOSU E844IJc2ItllkUJuGGSZ4b0GpZJNTgfbm3DChiaP65UYpIRpWqhlmV+m6eF7ffr553L40TeongjO Z+ifDsbp6KMotglRa2uqyOWRIibYHo0mTniljJiOKaqUV45ZaKk3JihkqZ1S6iKb/pJWxOSrAaZn q5102olqpmZumWunPRKpY7DBDkulqpgWSuysS9YaK238NVtnjNSO+GuP3+lI4KeX8mktmoGe2euq F4IKJpkw3hftfus+y+K0JOnH0KkezjdevfeGae+D2cqHb40d1rkvqfguOujBxCmqacIFA6scvCPJ 6y670VGk7LMXN1rxxGxlLKnHHG8GcpMjh4wZpCinrPLKLL9psnMl9xfzy5TNfJzNNEeGc2875+xY z7kB7fNiQs9W9NCIHd2Z0kgXxrTITRu9G3rw8Rtf1c3lJ5x95smrqKGABpz1qgJbXfTTUQcGdJhU 4kqujxF+C6KwDp8prrduz+lb/tpLX4Qsr1IWm27bxBaOpLjDWulluRtl/LXBYkMO8NYQFxs52P02 Suh47Jn5r4KXa47tvu1VHi+jkOcnueaUM9yvvxC7p7Wgn6M+uMXoNgj3tZfa/Xagi+ctOOKINi4g 3W7Hzfi44Ib6acAaz80g9LkTn2zihEs/HPJkKg84t2+Pa269waML8tr2Ovwe73SWjjDwSbrPqLZZ 46nRxcZi66v6d9vPoKh7qhD4OJW863FvVMzLnUry57vAsQ9w3RLcsow0s7UpEEu72x8FM4jAwlkv gYCCoANldbzmCc9wvlJc8ahXPAGSLXrLsxarDIes8ektXh6U2w3xpqnbielv/qCa4a1w1zu8FZF/ 2RPTnD6IvRj+7X4lXNamRsin+rXNfxgkoAulCEMQSg+LNDzS+B6XInUpLILtsx1+4Me9IJLRhxJ5 4gcTeKopWa9/MdRf9Tj4xIw47nENVCIPE3XGLdqRgEKcIAL1FSpSGTGROASesfDHQd01Um9AzODI eqYws62va6YDW+emVx94gfJ1/2oQ6QY2Sig2RJIaPKIe88avb+FpgIb8YrIsucFKam8lDNQhCo3o yyUWc5I7hGPIPLZGYVJShzbMlLYQyaEcTTN8s4SmNpOIQxMq8JlXdCSw5hjIYQoSYwJaY9lCR0jV kTJy6RNgKj2XMFHCzp2L/vLTD1G3OfnZ7nQLIx0+2xm2UfbTjIQEXT/feMll8o0zaMtMRB/Kl4le xqIUzQtGK7PRjNqlo5MBqUfnIlJ9GnR+ZgOYl9a5unbKEaUlPedIGRNT7eUPm90zly69yEIg1QiY M7VMTa+pLEh+05k+VKEWwbifoHJ0gfr747xSGEtz+s9+TO2mUyVT0yoyM4qWK+QPqzW3AG5vq1yF 6p3SCVadIlN9rGOkOVmE1rSe1aa2qiPv3trIt24rmXStK2SGKjxYipGqxDxkL28JVME+pqsCzac7 /dlKrs3zRArVWi1FeVfH0nRoXfUsWUL7WdGWNmekNS1YUqsY1qq2K65N/tprWwva2SYmtofBrW2v 0rLe+va3wHXZbg2jW9qcpbjDnQpyjWeW5SYXKs7143GfS5gBcc555akcIFkqnXLJz6SpJI/rGJrZ uF6uu4Cl7l2AaFjvZfJcXWIVUcunR+Vdc4n3JRzb7qbevuh0gn0s6gH5qthcVo/AORzrJctpzP4C 5r8mCjAmE3zUceJolxeWoAj5683y9TG9DiYpmA6V2LHtU6wTXp8i66thFYcwqdZFcYND7F9ZvriX WapbVRH8y7/+aq1DHPABZUpjufwXj1W9IIObmdge87LFJwycHIcM4iK/ZcqITWIwT8jkFWcTyeG8 IRsZTGEraxSz5bVn/u1inNDLkjFzrAMv+fB5yiermZ8sHWiVzdyxp0QXJn/ms1ACvZvpChovhCZh cw+N6OA6+tGQ/i2jJ03pSlv60pjOtKY3zelOe/rToA71piNN6lLHSdSP/csmUf0zVSOH1Y1JNFFW DWuiufSkLzUxm1GZL6wxlLOaPa/XCErFVgZbdei9NebYKbuvSaxzcHa2eeGKubDBztf/lJ2LU1fQ m9mxiTb+MPXsW9jBYXmnR06ykm3ZQRDl1243jbcl+XvuNpIbr/bOXiBlrRYw0/vLVWIfhLMpzFl+ +JFojPKS26orLh+OrHtsIU6vauMyp3vFQWwTh3GaSzxuHMcEl7h+/huq4YEXfOFsPSQuIyhEvXrL rQE/Z8a1DEe/7vmiIxx2NaV9wd518tc0nKIDoylzYQkdxCTuVeI2G2eSc/vAPP8l/37Oz5pTEOgQ zXka5xzQm/44jxHWcYevZXJBOU/dGFb6lvrERn1PsetFz2PLrT7IuUJt7fRlHpIv/nAxC3nseC92 3ys8xPvOOcrDJLrU3+1Lvp/8mCCHWYEB7kSyDxnIKNzyupHYZGobPOH/wzc2m/jv0Qee4Pkl5soj bnqOS010k/Vkt7EUWWBXtpQ9NSntmK1GYstcnQc1dpbCe3gXn/KNP4c28aFX++2KDWt2dr5CV/lP 2fB70K+uNW39/kJr7cuW+9n3fm5NTf7yM0v86E+/+tfP/va7//3wj7/850//+svf/PjPv4ntb+Tw T3U6fuZ//EcX3fd/rxSAADiAHyWAEAdWS1GACth/DTV98yM6VgJT0PZ0MtF24CR8T3c1yJdd1wZ8 o/M+axaBE8h5LZZ72RI94yZ4nRV6VJR053Zvh6V3+dZ65oOC3QJGZTdHakdNG+YZ1NaBmrd4CUhL R/ReIsSDjiRhXpZFuxNAFUhkZZSCDchUKGeAOzh0aRdfTmh5WzeBNVRWMBRKz1GEDhh0MnY+31Z1 cOVXWFd/k3NxB1d3jHWDNxcxapiFFod2p0dyTAiGYThxD6Rv/pokPnJHhFhYeH8HZUHmdRI0iF4U hmAGifGzWOLDNnsoEhzIcB3GY0klSOBGeSGngKETTx5ILoBkTxmIbcu1ME4HbNLnZtVHZ/kSI3hm gs93i/YHgQ0YjEkBjIXYbwwYZMKIFMRYjGixjJ73jEbhjMxYFvpXjeY3jW1xfUEhjdgoFtoIFNzY javlakkojoa2OtiWTldFfRRohLsYWcl2gLAHg+bIFfsVPi5XfKUoiVl0hPNld2JohfWoFVDYc5Oo hwbZhx42i37YeQM5juGmYAuGcHN1cGSGM4oHkA+JFQUpWZ5XhiOCdT/VZuIUR5OnkRtpFR3pOHQX TSw5kQpH/kQOl5JjsZJcyFMUKY89+Hh7g1Q06Y19BYP7iIikGIc8aZI++ZMQiYtwpo4JdWxQiVAe GZXJh5S6J5BKKRXf+BPhmJVUsZU+0ZVeqZXkqJNjuRXWmJaldpZs2ZbPcT5w6RW+6JZyGZevBJYc MZd0aY92OS942Th/uZcRo5dzOIfQpZeCyVsiuZiI6RT7l5hoyZh/1JhN8ZiQmRWWuW1mFJh+Y5iX qZVo+GxRqZiZ+ZnKFZqoWZqOiYameZqWKZqzR5qe2ZpM0S7Blm3x6JqsSZurKZq2aW2cGUe5yZug 6ZvDKTnBWRzHSZx+9pvOsnuzeRTPyZyHaZuRcpvRWRTX/kmdzWmd04mclBmN38mdtfmd2wmeyekm v0mew2gpwOlPy/mA7smeUSGWKImAN6FtU7l8rRN68YFrtRhtsAlr9pmehZaf80Zf98ZEC3l26MZ+ BfoV0TVzIbk/oFNucbdI7xehcomgRimEVIVLmAdfdyh+HApb+feRGIZFMGdIiBOCqoiVoXaiNUN3 QYhYWBU8XYaT7kejIQV51RSR/pl3Dlqi3uejP6qDqAdfrBdyI9qJnoakIfWOupdm0Ec2vNhmTWl7 6Sel9OlQ5biGXwqmZimmYzoxXnqm6BSmyaimsZKmbqpxx6hMcfoxcwqN5ZlVmwmCvKZn/3lZ8YN7 sIhm/vNEXumSOahEhrdUSlKYgXeGXbqGWfTUOoOqNnfKkEqReyXkXvo1kk5WpF14iuUkYDfIQjxm qgVUoa7XoiW5WZzqenoBpw9IjxWHj2YYpM0jbipIlALWQ9cVqh13kiCkeKoChBQncGJXljfZppUJ fMhXSb5DhRlWrDB5iaZHScw0iMfaKsvKTds6PMVGoacYq5dqXmp5aicik9s2LTk6XxkZUA9kSoln he8jN/Qyl/cadb1YYB24n6pZF/aJltPGeU/qY+4aVzu5pEyaTLtZcAajREbldz4VT3DHsNAKpVdW rgRJq4dIejiIq6rXiKEIemX3eUsorJFIlBC0dxUX/i6CEbBVUXwCiHlrNUDGWpJ0lKw7VLKHV5BD SVbkxEcBaZNSV2NsKqFcepe92JS5SE/LRj/C1o6xOZpJy1mGWntWC6gDOlBXyo6qZIuy2G2sBH5H W6dRA7Nmi6Yam7Yvg7Zs+6Zr+7Yc47ZySzJxW7drWqZhsbXTVqnV1lL6JKC59maSymsj2KfzqJ+P uk4p1bgkKG6cY5w+SDvIpnN6lrFli7Q3x4lRpTveA7Icxqo5aIOlaD41qISih3E+u3ruFpTdMypG uhZ0q1yKpqAKK0v/drPWOnCl56T2mqoLC2XfalTptm8sh4luMbtkKUVjI6LepUwuB2B5960ES6QX /no7hvc903s8ntpB2Pq7Uoa82Ygc57qWT6mc+lg/FvkwVrRzVVi0PmeCcWhY/vo5lEuyRjc7swi1 1vqpIraAzNuKn7qjZRa8AgMuhOKAuxJGH9pzIOm9uvislVhUCey694uxS0mAtTtytxuQClu68lWr 8Fu6Jpuwwvquv9OQB2nBOSmBALvBttqPaPSzdXe7R8hiMyazJYyzrAvCQ+q7hljAfPdx50hS4Wls V/q087jEXXuC78SUxFZnMggfVRt8XIt70YcwyHZnMFqlpgS2lfuvNYm3nmWgZKysZ+xUZpzGFcXG aLXGbnxmcRxUcNwSFnSVVeuoSRyohKts4uW3/n9bN4y7i8pHfWJrOorLvPK0f4lMgWHcvpebwQAM wwX8bdy0gjCJQaQbZo26V4hnvK/bpCC7dj0cffxKiFS2t2e2Myz6jMSaYzS3u8c7hGHHwY4YqnNn oTGXwikrcan3JSULW6tMqM3Eu8joraaoRVZ1sQ1HLQzKwLisvTTnzF00wdwLvg8XzHxZvtxsbj0p qsc8tfxLdP1KWdVMS9+VrOlcv3DozWfYx+68nyZMxOoVO9p1dBB7y5rVhA9rsYuYeYTahIlEZor0 pDvLyBDMvjoMiezaXH05WqCngm3XwiwbxI5HYSRMofRM0Os7djbHzx5szciKwYfpJkV8n69s/sO2 W3mnR7T0c3lNerqZ/LGT50a0vE0gfdEkHJkmTY2A6pSFDMhXLLhUeU+3F7bcNjBG7cWbo4FmJ3xU J3udXJXVJtRDTbXNx59HXJ1KO8dzK5lejaZgHdYYM9ZkrXGpmZ1nfTJpvdVrLVRtXcdvjZnHuZ5z vTR1HZ93rRnradd7zdfO6dd/DdfeKdiDXTPmOZ6HfVGJbdiLPVjzeX6PnXWZq8oKDE/3LMavuFBa vKVhJZWPHF5bLGXXBZv3ZLgfeC9XjT+l/cXrKCFy7RLK+5XLOqoWFn5Ml70/jLOJcoYr+zzznHoZ psmonLrwBrTghqr/E9tpWNlI26u+3WQY/snSUYhMPHurXrbAypyRCOl4L014FlywLbxed7ux0L1C GjRDEvO+nBi9BnbeLvpC5xyshyrIo6iErEunOLy/scrN3XzbX+hjDK3W/Uh1+YyhMEKF29XbqAug 89tXQifgaZZwBwyNzK0biPx2SHhModlgkujIRblFnBu+N+aSKvp17CaKErtxpDqjJmy66a2uuTy0 KMsrqzeiIVxyQDvSJSyz7q3bAs2xkwZhKQ3OvjHivLzD/uspe8R1I57cZIdgUK6yAO6/W/hpWCtQ FJvZe7PHGph8Wu7aENJrtZQ+SbxKVFtWGHh7SuxVmkqLDcOlAjzZyxsTkEXnMWvnUIXnHnmOmXz+ 54Ae6II+6IRe6IZ+6Iie6Iq+6Ize6LoREAA7 ------=_NextPart_000_0000_01C71122.93BDACB0 content-type: image/gif content-transfer-encoding: base64 content-location: file:///C:/Documents%20and%20Settings/Klaus%20Holtz/My%20Documents/My%20Web%20Sites/mysite3/THEORY/info11.gif R0lGODdhPQIlAfcAAAAAAIAAAACAAICAAAAAgIAAgACAgMDAwMDcwKbK8IAAAACAAICAAAAAgIAA gACAgICAgICA/4AgYP//wKDg4GAAgP+AgACAwMDA/wDP/2n//+D/4N2cs7OP7ipv+T+4zUiENpWM QY5eQqBiemJPrB0vvihmdgBFAEU+AWooE4U5akoyhQQEBAgICAwMDBERERYWFhwcHCIiIikpKTAw MF9fX1VVVU1NTUJCQjk5OQAHAA0AALeZgYSZtL29kH9/YGBgfwAOABsAACgAAAgJKwAdADkAAAAA mwAlAEkAABEROwAvAF0AABcXRQA6AEkRERwcUwAW/ysA/yEhbFkUFABRAEcaahkyZwBhAAAx/2EA /1MgexZDZy4u4iZZFlFGBGguSQdSj2oYuJAjFQBT/6MA/2pKEnUzbEpBmjdlC6QsFYMfsU4s/yBR tghkkm9WC1lDrTZyErAzFwChAHdfH4lHcbBDHLctfQCGlXpuIyafAHOpAQAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAADKAKxbASAdwpRScCSqTAqUiTZue0R1kP8AqABx/98A/1aRSjRI+MwyguRBcGjK ATa8QgCa/5Yit4V9MyW3jDZa7Vz/AP9IACKbokLPTcJYUiDTlaUk4HNWtampANBvPGefWInPC/+s AKcu/uJZf0zcZ/8Y/zp9/7HQGMf/AP/iAN+aPVaBn8ZDuq9xiziiydFTzv+aZUbK2/9N/8jpakze 4P+Y/9/AgunspfX2zf/Q/7GsWmORriJMZY1OP1BwcND////n/2lpaXd3d4aGhpaWlp2dnaSkpLKy ssvLy9fX193d3ePj4+rq6vHx8fj4+LLBZoC/eMbw8LKk//+z/9GOo8PcN6CeVHaucHiewYNkv6SD 09E/Mv99AER4IyRfYA4OLL4AAP8fADE5ANmFPgJ3hbDYgVYhHQAAMIjIs6B5AMDAwOpwgVHxaf// gJF0zf98/6L////78KCgpICAgP8AAAD/AP//AAAA//8A/wD//////ywAAAAAPQIlAQAI/gD/CRxI sKDBgwgTKlzIsKHDhxAjSpxIsaLFixgzatzIsaPHjyBDihxJsqTJkyhTqlzJsqXLlzBjypxJs6bN mzhz6tzJs6fPn0CDCh1KtKjRo0iTKl3KtKnTp1CjSp1KtarVq1izat3KtavXr2DDih1LtqzZs2jT ql3Ltq3bt3Djyp1Lt67du3jz6t3Lt6/fv4ADCx5MuLDhw4gTK17MuLHjx5AjS55MubLlywIBaM68 maBmAJ4Pfv4HmvPn0gNPdybdeTTr069Hq0Y9O6hs07gLomadevVr0qZde1bdu7Rs4rE3106+OyZy 18JT66atfHnw4tN3R//NvPj2zMNh/uf+PvE28t/WQ4cXqr25b+C6e6uHP5++9vnGpduv31wndf3g GeQef/IFSB+BBu5nIG0LAnigc/rdJ+CEEQbIIIDvvfdgghIe2N+AGNZX0X8Klrjhh0Pdx6CKInYo Gn4Iirghhz6B6OGE7YUIo0Id9vhifBQmCKGFMoLm3oUOngjeikHGaOKALkrYH0VS6nhjaDYCxeKS WMqo3pQ+OpgfkWSKWWNCLnIp3ZgzlglkgzpW6WaaMF1Ip5HxsSkkisAx+aaTSD6Ip5lrcpShlWkO muSZhapZYnvUaZgcoMKJF1t2kuI0ZXCHwqkopNeJZumVl3pnqnebumRpek86uieF/p8q2eR5k+5Y a6OGRtrjqK2y9yV8ugZ6JZi2Dluhm8BqiWaLqzpqo54xhilnssfWxOeeo+aXpazaCjljlNWiGu6i G0HrqYDZpvjrmCBmmSqp4P7pLZw9Pesktby9Om+0xS5K7L6qztoihtvy2a2sf4ZJapnXZjStrYmS y1OV7L4Zr770Luytwu/aFOjFc/Lrb7HCbvwrwC01HDGNRe44qMos07tyyR1HtJq5jw4cs3+ttcZp pKf+jGmG0RXtm3WV1jyT0cb1HKywq9IqdHhAr4dq0yQ6d7TTuaF3s9VUYwfs1uJBd17S3HXHmdgO c7fdcf8lbaTUmNVt99145633/t589+3334AHLvjghBdu+OGIJ6744pPNbbbj1y1H3tRqMz021j/P LTbSyuGmqEeQo0v059gSJ3eOXTep9uqrc/5y561Xl7KGs+nKo8FcTww7eqrf3DN2ZrN9W0pIsnlx yQkTuiuh1OoprXwkKi0R8hw33HzGIKsOL6BxQl+ggnGvBDWZdM77Nco0Rf/9yMz3KrP05bp8Mrkd K4yvvex/f2eXVmq0f57yIp2+rnW+FxUMY/ajUQKNRTyJ/c9LsWrTTbaUJNIVMF9PspO8RoK/I8HK SyJj2LLyp6aHEe1HbWsZAHn0LSAREF84CiAC5bc+DIaLdg3c1wOXBCVc1ah8/vxr1JY+dr2UYYqE oUIfrc4mJqd9p10eeqIHQVieDaLoXTRzYZAOtkWdPStbUpxO2MQosZAgT2OCMt6CMrUTn43QWBRT Hgwn95FtpfGDZYyZ8yrYvgEyUHtArKIKgyiqJ2rRYq6i2swQKSgddrFNZzSjAwd5rkr6Covp6tO6 5Agt+KWQhJ1EIUI41j/38XFtoFyYJxdixysy8pTjetS/ShlK7AlsgQvk4CSRaKfn5DF9hDSRLPFD RFTaUSStbJ8JBSazECJsZypaUTE/yTJSwnKGyhSl+45HoK8VMJqlJMkZ3fVIktVLgzszpf3WiRJQ Ac+JlaOb68QVtMeBrVJX/uOhtjpFzfTwynJdk1vQ0vYh3znudGSDni+hQ9CFHk0lk3tbRCWFuVLl JHSpEyg+AzqeC/KqJKt0S0hZeZmRMo4oJk1LSllImZWe1Fd6cSlLIyPTl9r0pjjNqU53ytOe+hQw tQuqUIdK1KIa9ahITepRWaLUpjr1qVCNalEVJ9WqOrWO2jRfVluoPa5u0KsQ3GpNBfnVWYa1q2al n1jXilaMPI1umaPjUEcZNaHSNahosmshV1qzvjLEryQFHVvLOtizEratiD0sRAur1sQa9rGNVSxk ydrGsYrTIoCd6V2x+FesOjaykwUraEer1c+eJK2llSxpRZva0KJWgg+x/iwHZYtMzHY2sHllFWzR RyXGtna1rw2ub1m7W0malrjCPW5yVftbynKqo1g7aO4KycqC2tNtcLvuKG/bW9y+cbO1+2Vxndtc 5A53ua49L1PVq1z2Mte8x7UZCvfIvEhG0obdy2bG0MVd8nb1q4qcq3fj194Cvxe9wB0uSN2bXgM3 +MBble8joRja/6KSkRReLRXFq9ntDlh0SP0ugSHs4ASXuLyvXfCJ4UviFj94ww6ZpblwiMYIX5iG w8GjEvs7PRBbNao2xuyPh0zkIjeVqUZOspKXzMbY7pW0mewwf0EM5aJ92MINsV5mwUvUK7uVwSZ2 cZhfzGHPihnFYEbz/oplitqC6ri4W0Yifufo5RFf08O51Wuc/ZdmFpNZzWcer5n/7OcxGxrQMM5y VjNc3kSzltHEdbRtZVjnuB4TwBdBMKIPXehNN/qya+6zpjtNaglbkXs3PKSITabfkGGaz5SW8tA2 tWdqgnbJnib1qGnrZMfiWteiVnCmE4rd6pRNanCtrjsfJ93o0q7JX471qsko61ePaLDUC3aot73Y xGab24TedZABx2QlDzqy3w60uMFdZsF6m9PrVnefB6dlHo8bzudObbrDrW15iw/b8O43v+MruHpX mre1dmu5F87wpyK54RCPuIARZ/BqJzrhmQZ4ruM98ECrWLH7Dji7/jd+FV4bt7s6s7h4MT5pkIvc 3zCvMKhdTvKRA5vgTjE5SPjaO3tjGeH59mrIax7zl+eQ5jcvOtFlnvOf8FzaeJ62tQVtasMWr6pG z/qndfnuNUpV60lneoyxdDYnGju7Bn2bqMheV6HVxnVq//mHWwnxqftXxojSeNjBnuKZWz3vXd/7 0jUMkeKZU4TqtKLh+4fO+93LelXH2J3vXXFD6X3xSOc43++988uHU+ibtzmb00lfSt7rjhorvZzT 2chrQ731Kmd1uwvv+f3qO/RKp3pHlot53uO+45Jm4QH/qEbYg3V3fqxW8cnJTJRP/tIX7znoJE79 6kPbjNbPft0z/j58SNkuic8VJfJb7/16SjSiwV/0qaM/d7v3OPBXD285NT/4rdcW/gST/yvpL/hI 077787N8qjYrFjSAAogsOMN6tPd60Ldy0rd7tWcfeDV/Avd77RSBb0VRn8d/HPhlAChntLZLa/OB sieCITh7PqZ9IeZumZd/upUzoleBp4WBV0MemAd8FlhFzIdGMGN7HFKAjAdIEBN1T5dyFWRU9taA rteCXjdxN3hr5lZqfoduiDJVzeRrUThq8uV9BpVRZCNQIzgwXFg1AEU5aBNI72eEpdNl7Zd+vcaE Eqh/vReBOGh/J/d37yOHgAeHdeh/vaZzSxhLLyFblbeGE7dq/krofFSYh0izh3jYf1o4hbfHiEx0 hXyYg7qHZ4CYhv0SMJbHgLM2gVKWiHc3X1Woh5b4iB2Ic/eXeUm1h75Xf1KoiOvlc932iWqYRlY4 iu5nMyr4i78GUcA4jEVGVQ94hFdleiYXizC4fxtoirIYiVznivo3hD6IdyimV4RGb8doiNr4fL24 gHCYbnOIfzEod50Hf6JojcI0icj1jbnGjaAYYOsoefYIa+O4iM2oijLoh62oivVoiUO3j/QDj7NI bt2oi2yYi6TIifpoVqL4hO6YgUrFaR8HkNUYkc9IgZb2ivG4c8/WdtU4drEVkhr1ggfXhryEhPOI giQ5jhrJ/pH5CGgDuYmyho3MgYrOaI7tmJMM9ZEs2INFFFLwk0X5lYCxp5Il6JO6lUxu+JL8aGA1 CYt0SHjTiJEviJPMKJFMWXzvlY7PND60+EZC2ZCch4jrV4MjaXz3aGsT6V5TKZBVaYdgiZVwpZU0 aD56ZpHi2FkQOXkl6WR/2ZZQOT3EuFTTd5iKCWTis5iOmYwtp2h7tTLl4Um0FmW8JXWFyZacKYTK GHR4KZVh1nsBSZB0CZrQyJSXB3nNSJGTKU9nmZQ+VF/hSG1+KX6AuZl9eY+JWIidGW2XSJOjeYpZ uZG1aWcPWZpayZrlqJop+FF994a4VZbHWUNk2XxmiY5o/vmZSvlMmaibodlicdmaymmcT4mPdvlR oHdjb9mEd2mQ0ambqUmb58me15k8hCmZtOiUDtiSmemQrRWFLEaaC8mV8XmHUGhk1hmeFFlXa3mg tkhdsyY810dXJWldYJg2kbeh/Al0/vmd+tmHq0ieOmmgsQmBylWeaMWcsBiTXcmKIcozjLKfaemh DFmdPjeiHUegjZiKVvmPwKWiZcWiAumih3iasikTNqmkCTlkNQqiOdqP8cajlWia/oig2biW78iT rumesKmdSTokU+Gbcrdn2Xmhj5mmLCmMatqmR1qZXKaeXXmZ8EmnWjqncSqnNMp+SUmmwCmi/dac 7Die/kcHqNwmqMR5lyeqN37Kp9vpkjEajYHqiFQJlEAqqYdKqZSonhDaN43an7xYn1emoy+HqHJp qVgKiZPqo5tqg4uaN59qo4/6nxuqqpkaoCXKl1eJiVOaoGWnpZ3KN26Kkhl3jjAXfz26k18pibZ6 rC5YpfSogWC6N7GKb925jFK6qklHqBdorDsKb0IqqnVTrSy3lNjqraWarqhal7wKl+AKrK+KN+Qa obxUR8N6r8kmTvi6r0uqGPOaktVaiugqnETHrTM4sNsapLkKo37zr32akG7ZrN9asLp6qRKrru+4 sMsaOA6rmfXKgobqrBS7rqiJsFspVkYqbBz7nNk3/q3S2a4OJqDSaLGkunQyq7IICbD0KqvIiam5 V7NXyq4+G7IQFK4/+jflaq0PC7Iw+7PZ+m8m+7ShKK0bm7NLe7UeG4hNS7SYeJFby6tGi6TUurNK m7VQqmz8OqwPl7Zuaow6+7ZhuptD+7Vze7BOG7UIS2/8WrJ327dca7d/e7F1e7Y/NaOBC7SIW4t+ S7eCW7hMkbh4u7iAy7iQm3uOqxSVe7hSW6iUu7lgd7lJkbmda3NeO7iiO7igixSn27irW7qs67mo m7pGwbb3ura0m6ay2xSru7v96Lq8G7nbmLsoBbuvm7fM+ruS23/CO7zAO7oed7zEi7zBu7y2Eb3W /ku60Nu8pjtvBbe3TLu92qu82Zu80luxSEu2SZu0Aku+1/u8uwq+7Iu9u/d2v8psIZkdscOpEwpd t2i26Uu266u54TuzqVq+8Du9n1gxlTqgh0dJ09St11VR5xPB/Hs585Mrt6umtpvBi8m5uVQ/8/nA moRq/WpA+jPCjpc1eKLCKNw0tCq3xRu/lju+rhW2MdyHQPo8griethRM4FLC2uRNqkZORAxASBmZ zivA7kuz52XDBnyQTLtETHOCFgNGcRdFagfEveMnypc87RJKL8O3SjzG5iu0uubE7YvAUZx8yKKP 4ONjFwypdhYr1lRNYsg/R6y1NpuFaSzHf/qu/uV2wEeLokH4czi5j/jjnUenRtPUyCPTS8NDyBN7 w4zru00MrwPMvez6wKpnx3rURxKpxXgkxEV0PQezRwoMQz2bsILcyibBgWicyQxbLl04gs8hT7fc bBTaUBhVy/9mOj4DzPuEV83GUJV4riJLya5sySmKyTKMwwWHvtKckpXJwY+5wdZMjG6LtVHXzfEa paLpyq1Lw8NpaX0si3q7r2Jczsrczswsnmr5xPwnj9xMeQAMoKzczuP8vvlskPqsyVbrv9Mcty87 snHImP/ssqvcqx0JmfK8zQINt2YbwOxsojJJxi+Mi5Psz+GJ0YSLGS48Hmv3gw/FNqUjrtPG/tBW 2tFJ3K3JzNGp+dBagyPTlTljtMXLVtN10sVT5Ig3WMcajbEWraweLcq416XBmdCDONLf5pUnosOO Zy08/aSfDE2fp3DZ3MGNmdWKudMxpMpN+l8HaNT6ychkx52psz0ZDZ7u2qpUa7CN+84PRqX6e43n 3ECHwkUiBnlUTMea4oJRZHtFzCwXtNBtLUvODNdAK9cV/ax13ZNP3E5jfX0d+lvu5DGnBNTYVGNo qMf97Niuatfhy9gEi9jFKdrPDLVRnceVfXxvjNk/WNWc+dNTttaR+tleqrEy3b9CDdrJqti9a7fO FkHdzNfY6YNL7TWEfaMmHdg4OqqHjaxV/jrUHk3aDOzWzbnbimswVCSUXmRJH/29ZTrQE43PKi3d zcKqSs25Bo3eG7XSd008GshsokNl9R1+xQZ+g+i9ktzbuf1N5lnJ5Fza7g3gqF3d0SzR3jzeicnV 2rzVDj6MEL3gU/e/4n3exmmq6+3SG72BKSvLFHdhP3lP+NtEUFI14S21Gp7dd03WzbriH57atn03 jQeE/KE52HJIjRfU7Z3hmrrhk9vj1xjjRT1sES7hS2jWOXbjGNTT5LPDxfrS4qfbLa64HU6ATtjS M+7HvO0fLUeGuKlQR6jjUI7EQu6cPKzdHO7faK6NwC2/t/3XhpuGftLXOVbYnXKA/X3m/nOt5SlO 0QSOshpb5GZ+UU735Q1i52Ke4yez44Ydzrgt439u3tEt6L9d5YU+NBUMG5JToVU8xVbTheHV6Y9N VkpeTryB5xTm6P5z5EjOpq6uglE+U6qn2VC54yCjwF65Q+9HwQN1JFFiZdqFzFcOz+L8zX/M540N 4mM5YQMYtMf9QiecizyY6Y5K4deOnmze58eu0Nqu7IHO7P7l7LbpalXHxUFYS6m2egXN4PXs7eBs 7JFO6F0+78su6S6lNOmteCV97iNeyDYOfoo6IvzN42F3s+Ie5AefZH6e75HqZgZIWfwZR9+9w0UY 0e/+3DeJ6fiO7LPe8ByP0voufcc0/lKUWcj4VfKDjPEsj+1cntIhD/KqTe80v/IT7d3x0tl9pEqP h58DKZjfV4hWBjZjM+17DuRIH6xHr+bdrvFwRmz0Tep0tEUmGeoelXbTxU+CpOfsCdTT3YkOE+u/ iM1iL3Ef7+VzLmHcPZuKbiG0veUbn/BJb+Uy3/Hw/vJrjvbdlcgovDHlw05wD/Nyz/SvHPNNL3b3 LKYTM2mJrO6v5Ez35dmET/jWPfnx3e5yXi+Mv3+Y2S/8lEvJPveDT/eHX/cif+hpv4DZ85tQ896g f/aib/dOL/mGH/vNPkGo31vv7Z3lZ+ITnDvEbvrCr/CWL/cwnPmL79kVTt4uT/uj/m/8A277fr6b ugxdaFf9rAPtio/Py6/gHp9nZV99ZB/+DTfrjpxfUU16d1/vaFq71AuroMx6rFntKJ33x5/t7l7e 7/8Yjk9nrw0Q/wD8E0jQ4MCDBhUiJMhQ4UOIESVOpFjRYcWIFzNi1Aix48OPCzGOJFnS5EmUKVWu ZNnS5UuYMWWaRHjRZsOFAHQe1Mmwp82fEh2GnPmSKEeBQHHu5Fmz4c2CTJP6TNo0YVGsWbVu5drV 61ewFJ2KJDsU51mqBctONBuW5VGLZ6teTXtV7VO0cseOVQvX7V/AgQUPJtyVr9y8dhUjJnsXZOPC SFfyheqz7d2Bez0mXhvZ82fQ/qFFDz7suK5jxqbtfrw8+vHkvFJL16WqmSfdzahd7+bd2/fvwz9l M+2Zc6hwq3hb+/XMXCha2xopJy7dl/Hp39m1b+fuFXthqMDfcrbduW/Q6uVVd2ff3v37kq0JS9fu fKN16rmxq8/fH/5/AAPcLjzS7hMPtuKmSqi4oKBrK7OdGowqwgbti4lBBTNE7qkKN5yKOAprYnDD BEcSLsERMWQrQuOOk/BEAWOMKzsLC4RtRrEkyxFHsCxD7DLaFAsSM8iGNBE/IzN7Dq8fi1xNRigN 9K1GwahcTMold9SSrb8edFIvKYf0ErIrs0xNzNxwew216qJ0U7fdrARMzg9P/rTzTjzz1FPPOQkE Us0mwWSTvvV4PPPJ68wUVE0635yyPkgdbe40Fytt0dJDQSoRzjSvtEwqNhUlsqxNJY3RvlLr/FTV GpmrFKgXQ5TQUFNtXO6+MT3VVD8yRSUUTgJ73Y+zWgV07rs/kUTJL1CRbSw6khot9kZi5Utt1EyJ BfbINRHtNlTdhuV0WviO1dXba6NVt1MhF7NQWnKXTevWTqc7t9D1rM0W3XvpDRTfeNtDdVt+5XQV y3+D1TFg7zLk0EVNO/Rx0xQfVvVhUL+lOER2OfRo41hnZZi9d4WaVeH41v0WYzJLHvllrBqFF2ZJ B+4o2TJVppW1Z1emlWag/qldNmiiReVyX3FTOtjnQlHesmioaVI6aqrHTfNXbNPlduF0cdba6KrD FnvsqePL+GKP0x6ao2ZNDjk5neXdc266686TbLzzzmrmhmkc72eme+14cL0LN1xqvxNXCS7GuUaY 8MMjlxxy0fg27O+IlcyozblY45y4mFe80+3RNy/ddDtJTx11GFX3MPPVPz5d9tgnl9tu3HPXfXfe 68b8yZA4xw/Xoy3P2Xbkkz9cYXGbd3dRbWEyXnnqq695ZecTzXr4ykIXHUQRw4dQYtebKvF8FtNG Ecbhxgf5PM2LP7v19Vu3/v7msK8TNw+d6n+pF31NaAYKzpe21y1Keatt/ohST5Ju46Pi3Yt7scFf BeejoF91T3vZA9f0rDYqDZaJKF6bFwKFtSvotUtfi8JaegJnQRhqJYPbsxegQhhCmRylhCg0j9VI CKgwsUtzO7xNTgwVvy9hTYAejOH9bgY603FKRbKTYvpmosPzIexkYjnZFk3TvlR171NP5GHHPJdF /T2tiWtcmxopJ8APxrFsRxMWoVaYsA2aUILWIaIZFYXEdoXrjUxko/IaBzjBvRCOc/TVa1A2wp4F yk8nfNZ3VLNCr0nSj24sZCe52DtQ6s57dAzkv4YHLhZG0pRK/OFcJpSzTOZrk6T0ZC1p+ThFxvGQ YXmfhlJUv1LZD30O/uti+IwDN/HJRkMOE+L8SNSsjOnQltO85Rt16ThrUvN3o6ymNm35RCu2qIpd +5zwvOkSQjLrnOecYTsjuKTgLXKdjLxQ3ObpSealsYPPo6E8J3M31tkPdq+7lEALSlDzGTShCsUY Q1mVqoYi1J/3dKI+s3VDR/KTkBTlaEdlaFFB/m9CaHRfHz16UpSGZnwhDWcKa0iejKZUpjO94JoA SSQOYrSHNOVpT7kCzmuhyG1VvKnHNupTpPJ0l91MZFOPl1SoRjVliHwqHJcqVaxm1SKh5KrditJV sIZVrGMla1nNela0atWaV8WlUydqT7U28ajfxGYu2ZrNk8w1rkXT/is+UQesKTIVliIzyl472dfD Xk2QSzlmIqFo0+lh0Zi+LOlCG3rMypoxfQoNpqzAeLZBvY2Z60uOMJ8pPipi6rKr1dIYYYW+ZELU rVDNJyX32Jmk1ZNH/nKgJpnkrEryK6iltFYLS3lc3qLrZkXEF86kyaQDOpeCgo1qbfc4Q9zyM4cL w6GRkmiemyTtuTZVZagmCUQhJXCn6M0jedf7x8HF0mlyRKp197XYDiJnvgMMokjC00cg/fcxloze cPH4vPMyNogFzG4ZHavYBjfSZ/IlHGILuVL6BFZ7/g3gBGfbxu9FM7Sq3R8PRWZScT7OtZdyL8v8 ayb/tcxiFZZw/n5Be1twLtC2uUyqO8lrXQ0G+KvcnWUPhVdOocLTaOo9pYsLHE/+ybjJrmQadhml RqdRuMqGDepjGyvgfhKVvojbkhIJ1uJcyRjM1WQyk1MZOCFrVMlCvGhMd7Zl4Zr5rTS9KwJJNNQ+ b7O/Ow4kbVC8z1XWC3reXbOeE01cIxPQgE17sMZgjGOwYTXQVb0mVRe31cmqjZloG+ZKRw1bUXP2 tGgjpqkxOBy4fVmZrKUf+KTjTGiGTLahHaxoRS1FLj8UraAccrBlZGGtIXuNm7ZqXT+cV2NDycKQ 5DKzO83JZ5M52sb6ynj3mmPAegqThHXyhYZ9bnSnW93rZjdZ/oPtYwjrEbIEVPa27R05+5a30HLe 8L397dF8C1eW7UXlvw1+z4Cn8Ef6HWjBD/5wbWJYP7Me6b4Jvl+IZ1yuiuWZgatVxDhrXOSJpSJg 00PxiIm53iNnebys/XJn/7PdM6d5zW1+c5zPLeIxb7anxzzVls9z5UDLuVevGHSE79znMPc5iJGu 9GkyHducjmyI1WdrUt8YooTV8BdXPWPobn2Yv05d/ViLWavE1rJZj3XaU8xquD+nlwVNO/v4S3IT b06EyBr3ru/OwEl7N74GnF+E3ayvVjY4ufPGTJsVeNzVOB7SOSKicftjzqlvXNKbj/ekI1/sLM2m 3081s+hf/pzGDAKXVMdLsA9Pr1wR6v2+LkTl0sCr6NH/nNMw7DiwuxZ7G/IY2nQ8Oa9Iv0lplr7O tF/LkeU9ZeiOK1lpHiKN7SwoatvZ8g7f89Bp1nvwh76/uZXeJy15ax1r0fjipyR2b2pHBWNJoOQG udXN93Y+Wn+nY8Rynr0IfG2rpfDjvBYrEobLNqBTPwc7pew7lxE6sfUDPOXLP/iyOMp7JyBisPeT MDBrwDcjvAlzOr8iJrqguFCrDZR7Kd3bGgVEr+W6JfjTt+Kqs+jLFAiSP88zoUYTnBekMoJzwd1L sO17JBHEuwLUN4GjPubSrRprrkj7rsnrNytzQtwDGCl74TTmeqnD8y2PWzws1CMVpDQeLMILiyKT 67KTI7CuQyfRkTXUqrUTex2uy7XSokOwI8G/UpszqkO1GzA71MOS6kM0Krm3mz+J0gtfW7tWc7TM iyGpo65rY8KnO7YuoacRbMRHzLS/m8T/mLZPo6ai852j48Qo6SuDIUW3CEVVXEVWbEVXBChUjEVZ nEVarEVbvEVczEVd3EVe7EVf/EVgDEZhHEZiLEZjPEZkTEZlXEZmbEZnfEZojEZpnEZqrEZrvEZs zEZt3EZu7EZv/EZwDEdxHEdyLEdzPEd0TEd1rKCAAAA7 ------=_NextPart_000_0000_01C71122.93BDACB0--