From: Subject: FinnPro: A Tool for the Simulation of Connectionist Models of Language Production Date: Wed, 14 Sep 2005 09:14:27 -0400 MIME-Version: 1.0 Content-Type: multipart/related; type="text/html"; boundary="----=_NextPart_000_0000_01C5B90C.B4B89930" X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.2180 This is a multi-part message in MIME format. ------=_NextPart_000_0000_01C5B90C.B4B89930 Content-Type: text/html; charset="Windows-1252" Content-Transfer-Encoding: quoted-printable Content-Location: http://www.uwasa.fi/stes/step96/step96/tikkala/ FinnPro: A Tool for the Simulation of Connectionist = Models of Language Production

Proceedings of STeP'96. Jarmo Alander, Timo Honkela and = Matti=20 Jakobsson (eds.),
Publications of the Finnish Artificial Intelligence = Society, pp. 79-88.

FinnPro: A Tool for the Simulation of Connectionist Models of = Language=20 Production

Anneli Tikkala
Department of Computer Science and Applied=20 Mathematics,
University of Kuopio,
P.O.Box 1627, 70211 Kuopio, = Finland

Hans-J=FCrgen Eikmeyer
Faculty of Linguistics and=20 Literature,
University of Bielefeld, Bielefeld, Germany

Matti Laine
Finnish Academy and Department of=20 Neurology,
University of Turku, Turku, Finland

Jussi Niemi
Department of General = Linguistics,
University of=20 Joensuu, Joensuu, Finland

Abstract

FinnPro is a tool for the simulation of local connectionist models of = language production. The tool defines an abstract network description = language=20 which is translated into PASCAL data-structures. The sequential order of = linguistic units such as phonemes, syllables, morphemes and words in = language=20 output is a challenge to any parallel neural network. For this purpose = FinnPro=20 uses a sequentialization machinery which allows for the formulation of = control=20 node chains/networks implementing linear constraints found in natural = language=20 utterances (Eikmeyer & Schade 1991; Schade 1992). The tool is here = used for=20 the simulation of a production model for Finnish nouns. Finnish is a=20 morphologically complex language where nouns can be marked for number, = case and=20 possession, as well as with several enclitic suffixes. The tool = simulates normal=20 language production as well as slips of the tongue, i.e., errors brought = about=20 by adding noise to the system. In this paper, the behavior of the system = in the=20 case of two simultaneously activated words is presented and the results = of the=20 simulation tests are compared to data from a corpus of Finnish = slips.

Introduction

Cognitive models are constructed to account for empirical data about = human=20 behavior. The models are made to perform in a manner that is similar to = human=20 performance in some respects. The purpose of such modelling is to obtain = information on how that behavior is produced (Bechtel & Abrahamsen = 1990).=20 Although connectionist models in many respects act quite neuron-like, = they are=20 not seen as models of the neural level. Rather, they are to be regarded = as=20 (neuro)psychological models.

Since the 1980s researchers have focused on connectionist models in = order to=20 model and simulate language production. Several kinds of connectionist=20 architectures have been proposed for modelling natural language = processing,=20 child language development and language disorders (Wright 1995). After = Dell=20 introduced his interactive activation model for language production = (Dell 1985,=20 1986, 1988), much research work has been done in that area (Stemberger = 1985;=20 McKay 1987; Eikmeyer & Schade 1991, 1993; Schade 1992).

In our recent papers (Tikkala 1995; Tikkala et al.; Tikkala, = submitted), we=20 have introduced an interactive activation modelling tool FinnPro. Its = basic=20 psycholinguistic assumptions follow those of the SAID model (Laine et = al. 1994;=20 Niemi et al. 1994).

Fig. 1. The structure of FinnPro

FinnPro: an overall picture of the tool

FinnPro is a tool for simulating the retrieval of Finnish nounforms.=20 Information about several nouns in different forms may be included in a = FinnPro=20 model. In the present version of our tool, each simulation run produces = one=20 wordform. In an undisturbed simulation, FinnPro outputs the correct = wordform;=20 the production of slips of the tongue may be simulated by adding noise = to the=20 system.

FinnPro (Fig. 1) has two kinds of input: (i) a textual network = description=20 for the creation of the network and (ii) information about the = simulation run to=20 be made (target word stem and form, and some parameter values). The = output of=20 FinnPro is the (correct or violated) wordform produced by the tool.

FinnPro is based on the principles of spreading activation = (McClelland &=20 Rumelhart 1981; Rumelhart & McClelland 1986). During the simulation = process=20 activation flows from the nodes representing the input of the network. = After a=20 few simulation steps, the most highly activated phoneme nodes are = selected as=20 output nodes. The correct sequence of the output phonemes (as well as = syllables)=20 is controlled by a sequentialization mechanism which follows the ideas = of a=20 corresponding tool used for German (Schade 1992; Eikmeyer & Schade = 1991,=20 1993). Furthermore, FinnPro includes a control structure to maintain = Finnish=20 vowel harmony (Tikkala, submitted).

The topology of FinnPro networks

A FinnPro network consists of two subnetworks. The selection = network=20 (Fig. 2) includes nodes labeled by the constituent parts (i.e., = morphemes,=20 syllables and phonemes) of the words and by word form specifications. = These=20 labels are mnemonic only. The control network (Fig. 3) serves for = sequentialization and maintenance of vowel harmony.

The selection network consists of hierarchically arranged layers of = nodes.=20 Within and between these layers, the nodes representing a linguistic = entity and=20 its constituent parts are linked by connections as presented in the = input=20 network description. Connections within a (sub)layer are bidirectional = and=20 inhibitory (lateral inhibition, not shown in Fig. 2). Connections = between nodes=20 in the different layers are excitatory and (mainly) bidirectional.

The control network consists of two control chains with a = hierarchical=20 organization. The hierarchically higher chain represents the structure = of a noun=20 and the lower chain the structure of a syllable. Furthermore, the = control=20 network includes control nodes for vowel harmony. The control = chains are=20 composed of two kinds of nodes: control and valve nodes. Each control = node=20 represents a linguistic constituent, i.e., a syllable of the output = wordform or=20 a phoneme of a syllable. The valve nodes (the black circles in Fig. 3) = control=20 the selection of the control nodes.

Each of the control nodes of the higher chain (i.e., nodes = representing a=20 certain syllable location, e.g., the first syllable) are connected to = the start=20 node of the lower chain (representing any syllable). The end node of the = lower=20 chain is linked back to the higher chain nodes. These links (not shown = in Fig.=20 3) enable activation flow between the control chains.

Fig. 2. Part of a FinnPro network lintu - mikro: the = selection=20 subnetwork.

The selection network and the control network are linked to each = other with=20 connections (the small numbered circles in Figs. 2 and 3) to influence = each=20 other's behavior. Each of the control nodes, when activated, sends an = activation=20 jolt to all the nodes in its control area, e.g. to all possible first = syllables,=20 or to all coda phonemes. The selection nodes, in turn, send activation = to the=20 valve nodes to direct the appropriate route selections in the control = chains.=20 The nodes representing vowels in the selection chains activate the = corresponding=20 front and back vowel harmony control nodes. The vowel harmony nodes, in = turn,=20 send back excitatory and inhibitory activation jolts to the vowel nodes = along=20 the linguistic vowel harmony constraints of Finnish.

The dynamics of FinnPro

The simulation of word production proceeds in discrete time steps. At = each=20 step, new activation values are computed for each of the nodes. This = makes the=20 activation spreading process a parallel phenomenon. The amount of steps = needed=20 for the production of one word varies as a function of the number of the = syllables and the phonemes in a wordform, as well as the number of steps = used to=20 process each of them (this is a parameter of the system). In our tests,=20 simulations take less than a hundred steps.

Fig.3. Part of a FinnPro network lintu - mikro: the = control=20 subnetwork

The simulation begins by activating the start node in the highest = control=20 chain (i.e., noun), and ends when the activation reaches the end of this = chain.=20 In the simulations, usually, one stem node representing the target noun = and one=20 form specification node representing the target form are initialized to = full=20 activation (i.e., 1). The initial activation values of other nodes are = zero.

The activation values of the selection nodes are computed = using=20 formulae (1)-(3). Value n(i,t) is the activation change of = node=20 i in time step t+1. In formula (1) = a(j,t) and=20 a(k,t) are the activation values (greater than possible = thresholds)=20 of nodes j and k in time step t. = Symbols=20 j and k represent selection, control or vowel = harmony=20 nodes connected to node i excitatorily and inhibitorily and = and = = represent the=20 excitatory and inhibitory connection strengths between nodes = j and=20 i, or k and i, correspondingly. New = activation=20 values are calculated using formulae (2) and (3).

	(1)
	(2)
	(3)

The activation values of the valve nodes and the vowel = harmony=20 nodes are calculated using formulae (1)-(3), symbols j = and=20 k represent selection nodes and (in case of the latter nodes, = also)=20 vowel harmony nodes. A control node is fully activated when = selected.

Every mth step (m is a parameter) is a = selection step.=20 During that step the following algorithm is performed:

Compute new activation values to all selection nodes, valve nodes = and vowel=20 harmony nodes.
If one of the phoneme or syllable control nodes is activated, find = the=20 winning selection node and, in the case of phonemes, output its = label.
Select the route forward in the control chain and let activation = flow along=20 that route one node further.

In the process of producing the output string phoneme by phoneme, any = winner=20 phoneme is inhibited after its selection. Further, any winner syllable = is=20 inhibited after its selection. These inhibitions make it possible for = other=20 constituents to proceed with high activation values.

When the network is disturbed by noise, after using formulae (1) to = (3) to=20 compute the activation values for the selection nodes, calculations of = formulae=20 (4) and (5) are performed. Noise is a random value taken from = normal=20 distribution with the standard deviation value given as a parameter.

	(4)
	(5)

Simulation tests

In normal simulations one of the word stems in the model is activated = for=20 each simulation run. The tool is capable of producing correctly each of = the=20 forms defined in the model. To simulate slips of the tongue, this = simulation=20 process may be disturbed by noise.

In our present tests, we chose the most robust phonological frame = constraint=20 in speech errors, the initialness effect, for closer scrutiny. The = initialness=20 effect means that in slips of the tongue, particularly the initial = segments are=20 prone to errors (e.g., mait a winute < wait a minute).

We sought to test whether the behavior of FinnPro would follow the = lines=20 suggested by Eikmeyer & Schade (1993). They suggested that the = initialness=20 effect could rise from the dynamics of the production model without the=20 assumption of a special word-onset consonant slot. Specifically, = Eikmeyer &=20 Schade proposed that at the moment of the selection of the initial = segment, the=20 activation coming from the target word node is still on the rise. = Relatively=20 weak activation together with some noise in the network would thus = increase=20 chances for the mis-selection of the initial consonant of a competing = word. In=20 non-initial positions, chances for mis-selection would be lower as by = the time=20 they are selected, the target word is giving stronger activation and the = competing word(s) and its segments are dampened through lateral = inhibition.

For the purposes of the present analysis, we devised FinnPro networks = with=20 two competing words: lintu 'bird' - mikro 'micro' and = lintu=20 - metso 'capercaillie' in inessive case form. In both cases the = models=20 included two first and two second syllables and one third syllable. At = the=20 beginning of a simulation run, instead of only one stem, both word stems = were=20 getting activation. Each member of the two word pairs served as a target = in the=20 simulation runs, and the to-be-produced form was either in inessive case = (e.g.,=20 linnu+ssa 'bird+in') or in inessive case+possessive (2nd = person)=20 (e.g., linnu+ssa+si 'bird+in+your').

The initial activation value of the target word was 0.55 and the = competing=20 word was activated to the value 0.45. The number of steps between the = selection=20 phases was 5. The standard deviation of noise was 0.01.

The simulation results are given in Table 1. Error rates varied = between 104=20 (10.4%) and 163 (16.3%) in the 1000 runs for each of the targets. The = error=20 patterns show a strong preponderance on word-initial consonant slips, as = was=20 observed in the empirical data (an unpublished compilation of Finnish = slip=20 corpora: 255 phonological slips with 21 onset-to-onset exchange errors = and no=20 coda-to-coda errors in word-initial syllables and 10 consonant exchanges = in the=20 non-initial syllables). This effect is due to the relatively lower = activation=20 boost that the initial consonant is receiving, as suggested by Eikmeyer = &=20 Schade (1993). Other types of slips were also observed, including coda = errors=20 that were not observed in normals' exchanges. However, their rate was = very low=20 in the simulations.

Examples of production errors: in the test with the network lintu = -=20 metso the following wordforms were produced with target wordform=20 linnussa (compare to Table 1).

minnussa	92	word-initial consonant
metnussa	2	whole initial syllable
mennussa	2	word-initial consonant+vowel
lennussa	2	initial syllable nucleus
metnunsa	1	other error
letnussa	1	other error

Table 1. Simulation of the initialness effect. Each error = distribution=20 is based on the first 100 errors produced.

The activated word pair	Target wordform	Type of exchange of errors	Number
LINTU - MIKRO	LINNU+SSA	word-initial consonant whole initial syllable initial = syllable=20 coda other error	96 2 1 1
MIKRO+SSA	word-initial consonant whole initial syllable initial = syllable=20 coda other error	97 1 1 1
LINTU - METSO	LINNU+SSA	word-initial consonant whole initial syllable word-initial = consonant+vowel initial syllable nucleus other error	92 2 2 2 2
METSO+SSA	word-initial consonant word-initial = consonant+vowel initial=20 syllable nucleus+coda	92 6 2
LINTU - MIKRO	LINNU+SSA+SI	word-initial consonant whole initial syllable initial = syllable=20 coda other error	95 1 1 2
MIKRO+SSA+SI	word-initial consonant whole initial syllable initial = syllable=20 coda other error	94 3 2 1
LINTU - METSO	LINNU+SSA+SI	word-initial consonant whole initial syllable word-initial = consonant+vowel initial syllable nucleus initial syllable=20 nucleus+coda other error	86 1 5 5 1 2
METSO+SSA+SI	word-initial consonant whole initial syllable word-initial = consonant+vowel initial syllable nucleus initial syllable=20 nucleus+coda other error	86 2 5 4 2 1

Conclusions and future work

A FinnPro model is the first proposal of a spreading activation based = word=20 production model seriously dealing with the morphological structure of a = natural=20 language which is quite complex. Our simulations so far have shown the = ability=20 of our tool to produce error patterns near to those of normals both in = the vowel=20 harmony behavior (Tikkala, submitted) and the initialness effect. In the = future,=20 these simulations will allow for the formulation of hypotheses = concerning more=20 details of the word production process.

References

Bechtel, W. & Abrahamsen, A. (1990). Connectionism and the = mind. An=20 introduction to parallel processing in networks. Oxford: Basil=20 Blackwell.

Dell, G.S. (1985). Positive feedback in hierarchical connectionist = models.=20 Cognitive Science, 9, 3-23.

Dell, G.S. (1986). A spreading activation theory of retrieval in = language=20 production. Psychological Review, 93, 283-321.

Dell, G.S. (1988). The retrieval of phonological forms in production: = Tests=20 of predictions from a connectionist model. Journal of Memory and=20 Language, 27, 124-142.

Eikmeyer, H.-J. & Schade, U. (1991). Sequentialization in = connectionist=20 language-production models. Cognitive Systems, 3-2, 128-138.

Eikmeyer, H.-J. & Schade, U. (1993). The role of computer = simulation in=20 neurolinguistics. Nordic Journal of Linguistics, 16, 153-169.

Laine, M., Niemi, J., Koivuselk=E4-Sallinen, P. & Hy=F6n=E4, J. = (1994). A=20 neurolinguistic analysis of morphological deficits in a Finnish-Swedish=20 bilingual aphasic. Clinical Linguistics & Phonetics, 8, = 177-200.

McClelland, J.L. & Rumelhart, D.E. (1981). An interactive = activation=20 model of context effects in letter perception: Part 1. An account of = basic=20 findings. Psychological Review, 88. 375-407.

McKay, D.G. (1987). The Organization of Perception and Action. A = Theory=20 for Language and Other Cognitive Skills. New York, = Springer-Verlag.

Niemi, J., Laine, M. & Tuominen, J. (1994). Cognitive morphology = in=20 Finnish: foundations of a new model. Language and Cognitive = Processes, 9,=20 423-446.

Rumelhart, D.E. & McClelland, J.L. (1986). Parallel = Distributed=20 Processing: Explorations in the Microstucture of Cognition, Vol 1:=20 Foundations. Cambridge, MA: MIT Press.

Schade, U. (1992). Konnektionismus: zur Modellierung der=20 Sprachproduction. Opladen: Westdeutscher Verlag. Germany.

Stemberger, J.P. (1985). An interactive activation model of language=20 production. In Ellis, A.W. (Ed.) Progress in the Psychology of = Language.=20 London: Erlbaum. Vol 1, 143-186.

Tikkala, A. (1995). MORPHO: A connectionist language production tool = for=20 Finnish Nouns. Report A/1995/1. Dept. of Computer Science and Applied=20 Mathematics. University of Kuopio. Finland.

Tikkala, A. (submitted to Neural Computer & Applications). A=20 connectionist word production tool for Finnish nouns with a model for = vowel=20 harmony restrictions.

Tikkala, A., Eikmeyer, H.-J., Niemi, J. & Laine. M. (in = preparation) On=20 the Production of Finnish Nouns: A Psycholinguistically Motivated = Connectionist=20 Model.

Wright, J.F. (1995) Connectionist Architectures for Language = Disorder=20 Simulation. Dept. of Mathematical and Computing Sciences, University = of=20 Surrey, Guilford, UK

------=_NextPart_000_0000_01C5B90C.B4B89930 Content-Type: image/gif Content-Transfer-Encoding: base64 Content-Location: http://www.uwasa.fi/stes/step96/step96/tikkala/fig1.gif R0lGODdh1wFrAoAAAAAAAP///ywAAAAA1wFrAgAC/oyPqcvtD6OctNqLs968+w+G4kiW5omm6sq2 7gvH8kzX9o3n+s73/g8MCofEovGITCqXzKbzCY1Kp9Sq9YrNarfcrvcLDovH5LL5jE6r12wYoAaI v9v0ejlOk8vt/D4XL7MXIOhXaAgFGJM4OHfo+LijtyiZKLkASKg3MdkI6fnpRtkoakCaIPpGuTm3 COr6akLI2Mk52pqJ2/lgCdvrC9LKiFCpOzuse1vMoPnb7Jwhe2krfXyKrKyg+rzNDaFtfYAqmwy+ iznana4+yxxuK14d725eHrx+78tbOu19Xd4QrB2+gc44zev3bx89a/YIOiwUkNa7bBITCnMQzdjD /o2Hxhl0N1GeQpHLJArkiLIOKobn9B18ORLjyowpa6b5dowYTXIkAQrEZjOo0KFEixo9ijSp0qVM mzp9CjWq1KlUq1rVcfKqVlAut3qF1PCrWIhAx5ptY2rfzlTn1IIsezbuj7Rp37a0hFOuXiHJQoJk +Pbv3sFAagGGGTPxRcKMI1Wk2JMn4saU8zxOGNFi2MqcWXzEXEyy4s4p4MYyre6zyMwkN5MW4fp0 TdUwWSOO/foDbhK7uxnOGdlfz9y8UReffc3vaNGLiVNjV9Jj3VKQc9rDI4h56tB5myvW7nxed7or ga8ez4v2uutZvXsHHx66W8EaY/b9l+vwy97x/s9GI0YRP4mlIs9/6rXV32u2sVfbNACO1lxm/CVo 1oLiCLgYdgIyOFlXFFJmITzfsSXeZecFSNOHg1mIUDUAqhchdxCquBeLC5WI42QxVucejXqFyGNA g5wypHn0uaeTj4zZVp9HAepHn4E89qhkhTJaZ2B+RmLJ5IhV1nglltJMcolPa01Z35dqrslmm26+ CWeccs5JZ5123olnnrLBYdwwFr054VGBHkcBfGwOWhSifOpJJVSKWsbooxwxE9Fa7bHjT3aPtWNK paxdehMrKaKXJqnyqcXdTzjlpeE7KcIiol19eViXNgbZiitLUnanRqtZsppeeSVqUit5+giL/qmo 08Gqk0lR/jZfk6Jx+mCDylHbJxjYtgbtiCa5GpJ03DobbZqNepJlhwgWKBmT6b6nnLTCoXVfj7Rx WC6Soj7p5WGG/oIvh6+eq+9w69qnTLU6ooFvScM9t5yMDSk88b5bbhPwhbKiyTGMB2eYMIYzntGw mAvbFRxkF1K8csSgupJxrMmaCHGDz7Ubsroq5Syrkyi61KUwK2M4NITIMgsxf8fC5S7POCcdJhsx s/yzzX9qGAGMf3IJMM8EOzxyxAZbvJ/XQYd6s9fw1hw01i2GDffJ6CZ35EhAsjullP2aiwvHoc5b Kbf8ig3OmQMKPO/Dn+RCrrxbQtt0kns7/q4zHduWqyqUmV+tLGhQgvz5wHMTm9xJR8+c8rDSfTzd 2QyzZboqB8pOc8Goxs745lyKvrimVf+MTVdNo4qiy21n+8VdY1bLXt+En9eeqrFHJ6lX1WOMPKNL XP8M99oH4X0z4X/vw/j5ZE/+Eeb3sn767r8Pf/zyz09//fbfj3/++u/Pf//+/w/AAApwgAQsoAEP iMAEKnCBDBxL+xroQPRBsAMPLIgEJ7iBCorvghiEBgcJosEO+k1QHxRhBUJ4PhOuAIXsK6EKJcBC pL0QBTF8RQ1FeEOuuHCGN1pKDjH4w7nx8ARBBMsOhwg2phSxgUt0RBMX+ERDRDGBU/RD/hUPeEU+ ZLGAW9wZEkvQRcsd8YtbIyEZRxBGep0RNmNMRxoD+MY1xPF/c0TbGoHRRt/k8Yx1fN0d8eiUPvJP kGYgpP4MSQZE4k+RYmCk/RyprT2SEZLJk+QXKekFTMpPk3+wJBI5uQVQvk+UWSBl+kx5BVR+T5VV YGWkPLnBP+oGll2TpQdcOQVcgm9ovOylL38JzGAKc5jELKYxj4nMZCpzmcxspjOHiYFnSnOa1Kym Na+JzWxqc5vclCAtU9gUXaoAUd9sYSDLiQVynnOdMPMmO5WIzlS6M5zxbKcN5wnPd/YumvpUijhL g08f1lOH9+QnPftpRIPm86D2vMBA/vfJUII6FKFI+ScNA+rPhwqxoRbQaEIjCtGOUtSMHD3hSI1i USJiNCkpbUJLwbjSinr0ES8llEhBmtGCThSnLJ0pImJK0oVK9KZCFahOiWpUntIUqCj1aUec6gR1 KjWoITXpVJt6VKvORRFQlRtRaopGpt6giFcEaxLMGgKpls+jZe2qFd26PbHagKxwRav64KoEtbKR eMv7Fq3Q5CGhsY6teLWDXWep0L2iZ2mtm9ZifedVlZ5UinLVwOWi5TN4AU5ydVPYCw5bBNBSsLIe 3OzgVEOgstGtck8U7RBcm0HSRjNMGcMb89QmIajCdpdZLRRMg6Sxw+UIdHlLVRR3/luYwiJBr2ml bXDX9qLa7UhXTkXuXJR718Q2F7hvU0h0x5a61mJXjSX1rU3L1l3BDtdetfOsC6y71t5u4rf4sZGf kFGkzs6ubi2Abw/8e0LZOjRqWnLZ5yhXubjtqagbRep2NQOsAufrdscrI0AnSxbtApJzlqJFmain Nm+FAsNv1bBMSdwH5vYUxVoUMHJYbFgXpwTAjpEvDGH84vLO96qJGm9oZYwSGmPFx0RQ8Yl5nOGd UiFx2UVyiZUshX/x1skpBvJriTxgHIvRxFHFcke97Ecd3/g0u5pV8HBHqdI5jXcLTupQtRpWoMlZ eMUK1u5UFthxgplke74ul0fr/peMFFhcgsPsrPirZy2T18GzNC2Hn9evd+UriELOQaWzZmVM99Bo ZWHOx4RrYclSucV/ju2mZxa96cWjYa67qKKllun0duy5qIYwp1P33j7fQdf/jfWpOdcPz/KE1VHz DK/HcGlZj3mvMhlhgiON22Kv8NiNpPaQS23Zpx34057u3OOknehRxxjb2TZS857lKlsbmslfO29O xZw1MuuOumpGM9lQB+zPWjuSNo63vNmNsuJ5696YqpmCw/pqOfoahPuuZL+9kfCHJPvX/mbwiuEN cXHn+M3mtfiROb5jj1O1wXB+t8ifSm6hTHxRGN9FxB2ycp+kPCgxn0HNDR5y/jfrnOQd3/nFQb5s rNqt4Vq4ubNdTkJ+sHlSRJ/rwt3YlqUz/eFIP3KrTF7VnjcVdkbvb9NZzugen+rnWc+5TKU+469D CsosbbdNuq64qmdU7UWneyCejg+7l1LvI2a7LcMN9Ir/fdpUx8jgvV54gBze2IlfxuIJ33LDP/7C kVe8EbuJ+WWGku/6nvnkTV15x39e1IHP+OjpW3q5n57ZqZf86hEe+g+/nvVlD/rsERv7bNye9jw3 ++45APfIiv73t+R8rj1PfGU7Ee9vPV1pebUe4yPe71375YB9OZDgH1z33YM+9EEMlO/HMvdEsqDS 2Pyywemx8bKv5fVN/+Xt/rCf++6fLcU1DfX5l3/88b+/8utfe4J3PueHG+l3YN2nf36CPcG0CnmG b/lHfgq4gNj3ZRR4D9rndvSXd3skM9knfYxHfRDYXJ+2gQkYDhLXRs6zERhoZPJXfFpzgR8IeWHn gsCnNysog4BXcjEIaDjngSZIHTD3QZeSg8kFhPnVO5mnhNBUd0dYhDxATpuBgaDXepYnhO/XbEH2 hKS3gzxYKLwzheUWgSeIgl9YgBsXgPD3gzDUG1sYX2MYhEJoGgYofCKYhqq3hmaSLWEohlU4fFdo JjTIcE44dcsTgoMIh0iIiCyBfNzAgsz3URfThYDoh+3HcOkGDTTnhqg3/oleSIeul3aEiILih2kO GH2i6IG01oDOd4qJuIl8YmZmmIV5eIeguB2dZn2x8YpgV4tWuB6laIGz6IWVqIF2KHP4538I6IrZ h4xqmETD2It/eIsy0Sd0hnbmtIyp2DIBporQ2Hu2F33duIrZUYbZSIkZ2EN8uAqoSIvmkokSWILm 2I7oKIzq2IzfKIDe6I5ZFncWxI76uH14uIt394+tqH4eFGr8F42WaJBHp3X0uJDjdojYQ42jlZDg RIz7l4RLyJG91EkFaYw+GHYDeXzyWIMVSYV1mJFbNpH+eI/8eJGJ2CuQ+GTBZoMMqZD46IyOuIcl VDI5GYmd6JJsOIRn/saTIEmRstiH/biSiyaUALiOS8mUOklZjRiRkqiUxYiRVwmP6weT9ieSMmlH ggiVWfmVMbl8SAmUX5iSU5mWJpmUs8VBYWGPj2iVT/WJfRUojVWWS3WXVSaO4EeKPWNnW0mVtsgs ngKMpRNsxhWQpAaX7YSLTCOFlTmZEOlFkemHlfmSwuc2RziTf5mZ/leA4YcaJOiXmhmNnPl/pfmY LNmUcdhCl4mS0VGbY+KWEhmbipiY22gORfObvjlhb7mbJKlSgamXg0lhvKJ9dtmS2MiNJDJ2wXma xmmRqll9FXg4c7gHrsmVTvmQcUmURwIqacaG3wlrorl81ZhZfKU5/ux5mDX5lH1piGPTLRnYdc5J lmuJm2WUZuB2jMSJnrzJk9wZMqZJhOxpnWBZnPN4gGFTnj2Zmg26gQgqjNT1jJ2pmwO6oCXZnxkq GFkxITWnn/MpRh2JosGYTh16k8+5Qalyjx2mkogGmxx6jvvINnQJWQFWldi5gDQKNaWImUPKbz4a l1KnnLcjl/IZn9IYg7ATnVHKoBvapDgZjnsZhYNyjWFQouF5pdopl1rKosnIpL73pBWoKKYYoDVa pVp5pmxJpL7CjVQalF4KdUungv2Xl6+ZSWMqlWZ6p+3WnjwaGOPJpnUKqF6pi/XSf4UqpOCJqOAo ggNjOI2KlSAa/mYU6o06uqZwKnC/CakTaqKzeYCsaamM+H+J5KcI6aKJiXOusaUEY4Dw1aWJyp/3 iamp6p62yae5tKpTaqf0qRF4ij678TLuVaSaepTNdn5nSY0WSqRRUKuSmp3POqeE2oCBKFrTmo/8 iZbfmpu76iIdyq07aZi8iq2euooHoaabp5aJl6ThOqO7I6e79q5ViKygmq5Ryop9eq+L011biqWx pSyx6qv/CpjdeaCqpq7nSbAFe1jlioeUpYq+eY0Da4PA6XBGSi+XaZmPebE7JKIGywQSi5hlqjL+ hrFRSbC36a4cq3DVWZ042qm66rAuu3cIq3CGuhAjijzVY6yb/miyvtij1toiaIe0PrmoQvurzkqt LSZopPMr0nWhNmuTgeivMNuxHSicIPuzShu1JJtXTXuqtkoWx+p9zdqwz9evK6qzN6qvN/uCYttl b3uJa1u15yqqwaqF6bcs9DaUNjoUGkuYHda2AtqmGpl2YesrsaikXAOaYzmq4ViRlMpZH7qbodmq J6mHM+udYpmpkxuSfaVpn5u5ksu3zNiYM0ugDplkyrq4lfuxwwmkKKe1V0prvERvfxupiGu2TOe3 pMi7pxu6qZt8cpu4XXm8wNq75rq8eOu7T/u8yNu8E7ueKYq9wxtlZLuvvzu9eRu93fq91Lu33ju+ YVm0onu+/vMaquK7vuBru/v5vq5bZeoZH9Y1tE4KKL+av1baJvjLvdDrvIfCvwHMsuqrIgBstzSi wLfLwAW8wAkMwQ4swTtLwR/SwLALJxksuHPCwclLhh48wRq8vxZMwm7ywdV7sig8wh0sJylcvtK7 JjAcvgM8wy0MwrIZJzQcvwiMwTiswkS7wUAcw+5LwCbswjtMxDVsvVUCWXT7BP3rpmqCIK4lxYpL wCQCxWPLrkbXglRMOqo6dDncujfcrm4rWAYMv00MxntqBVgDd19sxix6uHQqw1/ixvKUx5orv3js p/HavjY8wzPZnDS5e3vMxz58vlesvPPLuebryHprxJEc/rjGS8nQqciXXMSCrMky1MedfMGg3MOW XMGUp3X2aMqZ7Bws5LPgqon2Gz+tfBVy7D+ybBW0TBzA0sWYmG8DV28EF1jKQw95ijDimqyqDCar Ai4KO7LFRZh3FiO0wnWwujTcBchLZshfQWggw8s3aDXbvDcSdiqWK23NUrs5u7mVYV/d1ssX0SXD xqh8g4uvGi92hcukYbm3wWTD06nwnKNH95PuZc/ZrM3SrM/qdiK/c9DshWs+aF/sK60EXdByttAV bTXLuc8EB2pxF9AiE60uJdERpILs3NDmPDnuHG0AHWLEJa/bC8sgsiEZ3c7cJtMM/Wx+89C9GsUh PcvO/mXRa/NtkAbUkmY7mMqordaE6QzTjSNsGc3LKP3UQI1gQr01jHPTWfvJnYEsTc1hCssujbWj D3jVD0M7Y/2RSq3OdPbTzXNvtSAy5mky7exscF3Ux0zKohzE+ovXmHzXez3KkOzXf33Hgb3JbEzY ee2/h43YU6zYi43FjU3GagzZ5DvJJoTKIJjVE3TZM4jMWCTZb5jZELTZOtjXTPTZvfbSXHTaUPh0 Gms+yJnWVXHP1EkebATbS7LaNYbW9OtuEB1Bst3ahFWzsU0Vs73G6DWdUA1wwrU6uT1lU2Hcw23N aq3Ma/oN2hsVo53KpU2/Jl0qg2aaFyPOwE3eu727/pnlzTdYMRC2Wj1d3qGdnLrDz1/NtQCn3Zgt FdGdq/XlzPXlfOBx35yd3cGNjPON3DG60tDt3JZG4G9j4MWMvD+p4O/d2Z/74CydXhKe3wuOA/qN vgd+0SC+aURN4QMOyxZOYD5NveNd3Bw+Vg1+Ixeerw59zVhn4uY9lQLdXh4t3Wf8FAHeZtwNRS7u dKlNQEDOiUKuQEje24PdQUwOe/DNQFDOe4At2kTOi0pORVi+dlI+5CVu5aYN5k4ORFxuczw9SGZO kDi+5WNe2Tik5lxl5ANE5Q/m5Use5313523e4nMuQHW+YXuOQICOe2iavYeO6Imu6IvO6EsYpo0O /umRLumTTumUfryEPtkfnekDuumg2+mO/emeHOoNOeqjW+rKeOqmnurCuup83ere+uqRG+t2POuy XuuBfOtknOuHuuuc3utI/Ou+HuyJ5NvDrqpks8XGjs4zruyh2RLNnrAaDe1bq+nT3knJbu2hVO3Z vnl5vuve/u3cLu7jTu7lbu7nju7pru7rzu7t7u7vDu/xLu/zTu/1vtdVhOmf5Fb5PkT4Du6j5+/x Lszj6ld6COHBMSoJiprp3rj11jPx7dFm5imM6+PjflnyXM/eHTnkeblWDe+SFs9TLS3A5WjmQeLs XlvgdWvPQ81p8+HmnvI6gqw67lXbqOEMbzb2/s3jNA83Np/g6h7zK2/TRd3yecvvq5TzVb3zEc/j /Xjz6J7e7Y057X3UOKPi5wzzEhY4k6bx3h3N5gY52N7rDa/0wYwXfiWdDw+4GHr02jPwC+OAdM1X Ew9i1KPTzd72h533hL33gd33fv339i74g1/QqEr4Lh1whx/RLa34V8b4jX9qxWb2313F151qw+pt yY3yUh/NI4stuBIs0vyfZ4/I0B7ynEXiHr+d+kX1LA70HnYZRH36nyH7kr/cOD9cM08zxCbzO47Q 7v5dB5/QvY83/V1oxW76af+ZQy/Vw8/7v7/twf6Zy//8xC/iG73W/94/05/2zA/gsZ/S0B/4lHiS WifI+YOqWV/9noB1+6/P+SK/9WMX9uxG0gIf8suJ02c/Ls0s/tpfywQQPwAXutc8mFxURyKX9e3M s6aRLM0TTdWVbd0XjuWZru0bz/Wd7/0fGBQOiUXjEZlULplN5xMalU6pVesVm9VuuV3vFxwWj8ll 8xmdVq/ZbfcbHpfP6XX7HZ/X7/l9/x8wUHCQsNBwpgAAOw== ------=_NextPart_000_0000_01C5B90C.B4B89930 Content-Type: image/gif Content-Transfer-Encoding: base64 Content-Location: http://www.uwasa.fi/stes/step96/step96/tikkala/fig2.gif R0lGODdhMAJ3AoAAAAAAAP///ywAAAAAMAJ3AgAC/oyPqcvtD6OctNqLs968+w+G4kiW5omm6sq2 7gvH8kzX9o3n+s73/g8MCofEovGITCqXzKbzCY1Kp9Sq9YrNarfcrvcLDovH5LL5jE6r1+y2eweI ywEeecyuxbP1735vDhjHwVcimEBohfhimKPo90gTKLnhCIJYKYWpognDCfm5MqcgmuFJyRhmaqLK wgr6KkI66opAa2EbhWuJeqML+1vKeyg84Sth3IQ8SBzJDPwc6xzgCMhALdtQfSBJZ3D97Y2tEbig yIenPdydbi0+PWlOTMheK0h6qQwN7Q7BXS7sr92kd+TCdatXUB1BehW48Yp3kOBCcXYCqqPnD+K/ /oMWES40eFGfyGXSZnUEiXDgRXgq5zk0+TKYIXYatzl8eDOiQZUTKwJ0hy5mSj3n8o3cl1BgRIqo sDHFOfPn0qJQm+LCN1Unyp1Zu3ZNCdajWJsu11HVCZQRv6NsiwkdS1ZhXLlz5WKFu3WnSQyeauql +xHwXbwu39mVGvJwybaMH7wFVxeyxH6I88L125ATZsl+JU+uK5HqZ8Hy7BltzDahZ6LMWGvduNcy aMyZGcoezbk0x9ZWtbK2OXe17rWoi9c2mzMr2Zygs1UefVn3uJObee9+7Zp5889qr9/GPZy4cdSa oyYfm7z769joreP9flw0e+iRpXN9XBp4OP30/vNCpCPeeG2Vh5xb3kXQ13PCYbfYcduBt55rsME3 X1wSSjjhbUE1KKBIAW5o4IMZKsbegiPWod6IuTF4IILS4cSdew9m12GHmMhSnXLzYVUgab5BdWKI JUr1lY8q2vedacMY9l5ZR65XY2r8OOVkdECG9Rd9Z2GpIYcwJRYYV0NxmSNaV8JkZllFztgblFGO lN6Pb0Gn3VA0paNmUl266dhNK/nkFU/95alnoU8RqeegVb1p3HkbJcqjbT1VleKfSL5HwZyTAtjb pBXKpimkl9o5KqheMgqMoI9GaN1ia+HZqnMsniJpmPLZRttyUDKUK5eAbRcgqsIOS2yxxh6L/myy yi7LbLPOPgtttNJOS2211l6Lbbbabsttt95+C2644o5Lbrnmnotuuuquy2677r4Lb7zyzktvvfbe i2+++u7Lb7/+/gtwwAIPTHDBBh+McMIKL8xwww4/DHHEEk9MccUWX4xxxhpvzHHHHn8Mcsgij0xy ySafjHLKKq/McssuvwxzzDLPTHPNNt+Mc84678xzzz7/DHTQQg9NdNFGH4100kovzXTTTj8NddRS T0111VZfjXXWWm/Ndddefw122GKPTXbZZp+NdtrFxsl2226/DXfccs9Nd91234133nrvzXfffv8N uEyAD0544YYfjnjiii/OeOOOn6Z25JJP/k555ZZfjnnmmm/Oeeeefw566KKPTnrppp+Oeuqqr856 xMHy2fq+sG8Te8Fp3gcg7SCGWfu6onC05DS6075f7/B6Jabw3ig/UX/Gm7vUIfoBD3w5z/ueWEG/ kzr79d5/D3744o9Pfvnmn6802+gXkXcqd4uBt9WDX5Gei3VWEbjTJwkeLPv7898/IyQKgEkb4AgM OIRaRQNyi1DgLgI4MwcWAoK9oGAIJKgDDJJAgy/j4AQZuEELhhCEBxRhCUmoMRN+sHt3UOEKWdgJ FPJFhhVz4QlsuAoaTudULdChTGrmQ1plkIcxhGEojBjDCBKxiDgIYgeWeEMkZoN4F4Qi/sd0sQ5r LNAGtsji8prnvFtI8YMHHGEHx/jFKUaDi2M0yxfRQkVKNLGNSmIecno0Q5f5wou/s0ccg1EDKLpR eXCsIhvb+EY7Cq96zLuAFTH2SEKmMXeKXEYgEblIQt4jkvVgYykSuZtJ/jEzLdtjIpfnxUYCshl8 eWMqR4GiQ7ZSktSrpCpJWUo0SlKRqdRl8VjpSFemUZRPnOMss1i9Tq0yl2jko+7uKMRLSrGXycsH J3fhP19m7JpHnKM2u9mIb4LzjOJEgROPcU5HptNFCSwnJNeZKXhShpsndOcDfyBPheWzT/ZcYD9p 9U9D8oCeH/PgCPdpIILyTwgGPWhA/juGwBcqNAUNfSIOW3hRMWYUZRF9IEJJMlFZVfSQQezozjQ1 Q1Ut4X8p1YZJfcBSdb70Z4fDX+FWpQTDUa19X+Cp/ULaQLudDagr/ck8FzqFjx6NqEkoVZDiqVSB do6pAmSVkJCaDKrK76HZdA5UQcrVTWh1q2UoCQ+VMVM4hBVrY8Un7IhozZHKIKoFXCtMG7TE08i1 gaJrq1rRSUd/2tWSo/OrMdE5nXoONqWkM6w047lDhw50sWylLDAhG1nJHraxlp3rNLUpw73es3SO ZaJGLRrWtIbwdKV1gRXpqUPVYhN1rR2nOkdrzo1e1XS1za0vFepE0WKWtp2l6Dcn/npO4R41db2d IGpP2EPdfoq0xSWjRa3bCunyh7nVLeFszeha6dJ1p939bixzWF3lMol1zX2uVLEbXhu2V2vzTax3 fetZE463W4/rr3//C+AAC3jABC6wSgFo4AQreMEMbrCDH/w3j0J4whSusIUvjGHGMau+bsltdMuL Jq8+j8PLjWJ8w4kP75HYAW2F537BGMbVrbgdtv2wWmXL2btStrkvvs/3ZtzJ6M6gx7pSMYhpHN5m HJnG6sXcjOVZ3xe3aclZW3Fvo7pfRMWOxDymMoWErBQvTy3K7X0ymecJZP0d+cwoviw/08y0Lnty iOU9a5PjVchjgKG2WIbzlemo/t12dSdTey6ulcU8iyTuEM7QGjSINlkpnwSHyP9I8mZvbGmwGix3 nBqeKGuJSlBut53ZRfR6J7tje94ZXZxeLyV5+cyKgLKPpmbxYv1c61/lULEC4yOoYR3qWojaeqT2 8GSD4EOmrnpczkTHcoi5vV3G9Njo/QOjdR0L41LaWrnmq3PvSoTQ3nrZ7up2ptfo1q7et9TbpllI 2/3Ldv7zz/CGGXLNHTx1f2C+5I6zO6+NbWTD1sv9riuK8J3vI8QVxQBn2TUbnuicchLX9eYoaBFe 6aLqEuAFF9ojKx5xjS9TxxAn2WsxLhAnCBLlPvZaXlmecZUjseR/6pqdYR7z/qwakeY4pS8MeS7i J6gC6GGuMguJjmQomALnb2Z6DWcHcsok1VVOb3rVqK50pNvP1jaNOsWkofWgT31XWOi4vVlVda4n wqpZMHsuc65x9WVT7m13W8/eJvG5kbxuuYifu+k295tnGcfTZuiBRRpo1xkUx0GNpNsB69G8lzPx DTMKpZtM+eFWseog9Hq63g3kxGeen9UONwo9X67Tcznzo095FFke27CDK/Zrpn2qB1vvksqevy0u c+9T7ZhRyirk4K6xxwzbZXFC/ffF0PMw443pv6ipPqaJSsguccuvBlzbxVy+WD0yE+JBWvijHqKw fxm9YGdylLvXljzk+PxT/oMZzdLPPvHh2+zhMfKVcPfm+V39f6/WI0VRUO8HTYGRTLbmOAgiPQH4 U/Nzfsg0PVTEf/d3aXbESMCmSuknfykUIbu0fiFYgdxTU5CHEhiEOBGofq0GbcLHaOcQa2CyIZxi FQWoRdT0ayMoIq/XENKTgWL3backgPEmHwn3Vz1oXzYISzjoaSD4Iy30U3qRZ0nHg9nlVjt3CteH HU5YSTjyHCdmhdx3RPiWXO2XLWVYZ8A3hmKFcbqXdt6CfCDmhuzmW23YeoznMMn2UXfYWuLGc6wX KmbYLIC4ekSVTySEeqI3ZWYSZqi3NqAXepCoZMrXeq/3cK9CJbrmiMOi/lyP11LHVYmA9XGeiF5Y 5EDa0yIWGDB42BNXaIqsWE8aBYveJHg7xxJ8IohvsnihKFh4F2S06Is7WFVxA1Cd1g8Mo3cKR4z9 R1Jx4noy1zb7lhaMiIx2l25gQhJNhYu5OIbUEUfciCoEAgRnwox9oo30Bwn304Dlt2ndw3E6onaa 5z8m+AytoYq2447XNmhbp33KSI+/8H73ODBD1237iFi45I//+AoA8VQEcypEVn0I2WFI8JCbOGRp UoMLc3PxdScdyE61cY79+AkJopFwZVnaoUIvF5IimY5v9YbG8nLjJifqGI/yaHoytZBm9ZKPmEfG ZY6ExZI3KYugMI07/kksV6FVWAeUNllsOEmUKHUwo1iF6wh9tyVG2thM4KhtpMhqG6dsDGKEVnmV FOmVRukzxvCVS7h9BzmWCSk4MnZxD2WPYQmSp6WMcak6lxhQc/mLI8eUhudevMVNQMWX7CiRfylw B8dakwd6WtSXfrlbpkeJWhkzBNWYH5mFPemWgQk6wHWZwceZiDmOqkaZb/de5lWOYlmXw7h5oXNv e4l2qCma1xh5UwWbt3l4kDmRm4lbmmOIcpmbuqmQ8gZevgmcLymVZAlflkNV7XdyKxmLl9Ocu3dy nLdWFrkySYmcXkmW11maSjiVvHmYQhmeatN7EpdZVcWGZikyzNed/kmob/iVNnGYdskpnqU3VCdZ n2WpnKWGNo7liI5nnRxpNoY2oPA5j2BoLxnGoA3qoA8KoQwqURFKoRVqoReKoX7DiRh1n7NpbRfJ nl+XhkwnoAfKoTYnhyR6cf2ZXzanZAmKoE2ZXyGah0tGdBNnoiBqdI9FnsI5mhVEoxpZQTDqo7TJ o+QFpD1apMVHUkg6pDK6pB/aZuUWkFv4f001pQKHl+pJZ4LGfvG3gUXVpYC5lMQ5puqSZ+EHab52 HWRYazR3iSzKcL6jFusHR7X0a+OIamQaTUR6pmgaamw6hDRoHkYaTnyamfH5p+SSgJQ0qPqnRtQW fYYapZK6p+cC/oPSBiiCyjsJdXqbh0OEGZw/2acT+lsEh4rfmQa1B5UximAuBDmO0paJ6lCsuId+ EmOzx6pfYgnS+IWrCVK8Op6aCV3YOKx0GBbYOS062avXVZPHapfPuJulKpt0iZneJmLKGi0uKY3n xWI/VKapCYSuGq4CKXWK9q1ZMjGLYpi34K3P6qElNnzQSq+Mda6zGoakmqv6VHNVOUvl+hDg6pdM ga8C+66WsXLoWn63s67BUynkWrCv5q9sSaz25woDd5oZiJTYqn6jpq3JMhyPGUzOqoIiG5kiOSW2 2K3V+nwJO3+nJLK8iKYsMhWDcLA/aJ/RakvtSoXUSqsLG1gv/vuDHrmvA7krNQuxh5kUowhojmmy CriyB3ulRHutPqkr1NiQUemwqTiyUtuypxq0xCSMPZu0+Kqxz/my7OcmH3uU/UonP8eyWFupcJdi VEm2Bvuzb5u3JgZLRtgm1YiNsJKPvVl0T5u1UGt1PEu1qumzj9K398qG6IiwqjoezBojzsey05OR FBuU4NcU+lqueJto8zCcfJutSIspDhkrJ+iO/vSMBFi6kps9dQqv6RmazwZ9g4usMmh906W6NClp oHtwuFof8boq+DG01lqxYEUdtAu56xmI/iqzjBpTXHm8lVCosfuAkiJrtTu3UHVgQeGKtVJHh7uu /+GM/ukM/sFYm40oq8ILhmtrjZOogHSHjxnhd5YYKMkoUzzhU+bajRjJdwF8J+/Ta9EIvpVIayJU eNe7qRb0P4V4RxA8qqh1iw1VwfyiW9pVP77KmAOXH0VEDhu8TyRMuQMiica2Rdm4woarvw3slApK uP9CiPhZrKLrwS5MoMW5hge1itp5m/IZw/hHZbTnn+WpL3PYmmIIrPhJZrjnnkIsO2qYuetmvN7K Zt9ntVo8xUcHptPKwo0LwL7bfMPle5hkf3e7xDisxvQydCQLx5/0xTo8xsH3c6HkWcUExlbctbqb L6zAqVdMem/WglcMyESRxkh2iBwiaQlYtWuMuM5WsG78/pAwe8de+7hDmMjiqrzE5oC7+WdlPEgh +LynGal5useUDLdcmMpD7LSarLLamyEeZJAynK6jXJE2fMpNKMh06sWs/Mhlq4HUBL/eC5rJS4WL 7Hy4vMmKG8ev3IR+/MeM7EwwO66264PihyFpvJFoxn8S27P0dsvCNGzXDMlqSX3k58zHQ58zrMvy KsWLi8Tv3MJMnMRR7KuRq7PvvMhQ7M/6rMF8+JreGZOplbIYNZ2/Z9BsyxiEeJyl+JY+3MbruUK7 67oHnML0fLuuPLy9YtH5HFTR+bt6+asW3NFOVbjF2IpLknwkjb3HOZjhG6T1KFcq1SqHvEfEa5jp 27+U/hJkaLdsGXy8SYJ4Jilfa2ql10ulF6vTb9uqSiGLAkwYzLGLN1INwUuCIJw8nhokiMJ4GIy+ qHu11vuIy2gn1CcnUW13kYLWwsq/3ty7pqKXlmK/UL1dPPKqPN2Idz19S13X/RrXE13U73vTxWPV +0uTD0iB2yjG7Cu5eGzOoqjX3jxJjL3Uj0nNI8ytWJzW7muMtatMw2sYSmmvxbxydbTZmSnWaj3K kcouMELH3MOPQFm3g011YLnOCdw/p825Y1u/n23MEzmFnyjPEws9clvcst0klBiz22goT0h8eiUo HbnPxnzavALby4TcrM2wU/t5233bHYGzzA20NOu//n9U28at29y7luarvLVYKN9on77xiUuL3jOt Buld3IXhK8l9rSH8K+y63O0t2Ttdx7793jOX3eqq3voasK4s4LkNh7gt4X573+86GM873hTe4AgZ kJ1M4AKJVmkK4s+63QN74XRKt2Erth3OuQAOuhu+4v/W2i6Yj58V3GA83stL3zeb4jPbfy7b4v6t 4Rzu4D8elpZZ49mX2csb2wz4yS5OttANxzuu4hYo5KZys7NStfpd3ivLkHVsijlexlPt5FReptjn y0GelaU8sDBezC5sVGAOdbPNuCEO5aB550wO0mSuq3eb5Zw8nihNxjFW0tpt2XHe2yW+6E9+j0Ze /rYnbAa8IejzeuB7TOmxqwlwDp8yMshLirHubc4f2Oepe9yADrazOOVs56koCOmd/tSm3soVvr05 7dreXYyqzm2Zfunuu1rYxtRpdegVC8NF6+ay3tMCeuu0DtcM3Qdg9+pHaK1xWOmNXt173uscO+1P B9oox+lfVq2Wt6Why+hQmK7ZbrTeC3HfLuXP3O7HfrKVqo+4uOw1qu522JDUXuidK8uTvO/mfu6P CzFWLdhz9uh+xevk/umYjO5cXO/vfsBx3nAJX+7ubuzwjrhV7Oj5GvACb+/WjXAUj+xR2/CiPvLY XvKm2/EVH/Eg/6NZi/CxafHSqvEIzvErv/EA/rPpR/vyqRnzkV2vJ+/kBZ+2D6/OUZna1Q7wIn6e NF/aGc/Hfm58QC/plZv0Ke/wndz0RB/MeP69hzbrc9yOXU/kCrt9ugf0sxl78a7til72/oJX8suk 4rr1Sg/1Xi/MOR+EmMv1NBz3Um/2Wg/FgG/adY/xPVzmRq+61v72RS/nYwXtoGjyM2/3e8/3fd8v 1Oz2eZz2kA/toVmG/H7zeY7z+Mj4jT/1j29XkV/lABrLS5/4ae/30lz5ow/inr/5mLlOdnaRXUv4 9xz2EJ/6iov7d59HLvb6bX/5ig8LGer8zw/90S/9ijj91W/914/9DJYql5z75ybYa9/9xm/7/uFv ysHP7PVyyIcf+GOsnepPyLBP9rs2t1UPJ8mP+cPvzEDs/imt/Auf9QQQH/NA29hFOWm1F2e9eUcA ukBq9MyolMDQWU+SbWOxTN/Mpm/dnFF/FxQOicUe8FfJGTHLhjMBDUo/SFWIygxkZVoFl8Tzjsll s9gCBp/V1se609a43ufqBv7Ko5p2/x+wiA4tKfDJrQ5Ga/AHsSpmbydya5KjMqrJ0XCTczNy7VLQ kbFrcVSTciY0DvWr1YPUsvW1s9bWdE5zVQiOK5b3NPjQ73I3jbb0+HaZ2e5TF9mo9xRXCfE30aw4 Ovcmz7g5XPyQdrrWHIbb+tgKO7tsewyc/hJn3D6KTr2zElR/CtXdHSb9Gk2YRwiHP0XeXh28N3DB lodXZCnbN4vaQGg+Ai5QmAnWx0KSEk60hyUVlhqD8o1YqdLhO5AWPWG0JnIYTY/Q2KiLiY9XN5Ph VkBAeTQiUon0HiyVFg1dzZkLRRHq+EibT5yuhpTbOnQKpqZKJRZlYPRN0bREriajauhbm68C162d Kk/fT5l68IBtBilp4LFijSL5Gg8hsYY8n/JoCxRe3rlMgQn1awtw01Sal6pFW9aF2q7c4nLaQ7Ax jcf4JudkVXW05cuzxyHWCeg0wNZxbXtZ9VPvJL20iZ9YzTWxs8XJGaoOGdn44cnCWxe3/o5Ha0m4 XtmlTqe1Ot3X/6ZDvX5e6nPm0LXXZRvsOHLf/hwO5xcefX632O/+ebb+iHT26s+78RjaDSr89FvQ owAJ7Em2t8i7AiZWFNwlPoogqojBDkWRrD3/kIkqNhkgyW4+PUQajrMIPXwxuhgfzIq/7t4z0RXS LlQIHBZbDBHGIA0c8jYIXdSwRNb2A/JGGZ0skEkhpSzySIMUHGnGJXMUzcoTDUNRIxXp25HIKc2k 8Ekq2atSy7TMusqGjnozkUs12QSwuTLP3FM8PSWksUY160zJyp3+zPJNs/C0k9F/1OPzrzfDTLPR Se/MUb42E3lsTkNZY4zSKJvkENJb/hIdVBIeR4QvQ9dExXJAsegkMcos5BDTz1FJLdU0UlBVMc9X U0rUwUDdixUyOoUtVFZYNb0UyUd+hZZZXmv6cqtWXWUOVW2RPTTaHGJp579dKYv22WWdHRfEHqzt tdpgQwXXVh1zdSvOiPrEKqEEm90sWWfvRVeK3+Z9lww4cZqHunTPzRLcgh5GaV83jQ2RBYo9dTgT rzJ2I5TfrkR423EdlZdAuS7uWLsl2CWHOn/f0bhBgR+16pqbzSU54Yz1rTm6FWXu82VqbS5E3J/3 Cm3o9sR1gor6FvtYyX8hrpTnyj5Q+uGB1V0y3x8b9hq5sJ36EeD7ZDZbFXK1xnlr/rHJDiPrNTcG ulhgd01a7FV1rutpVaoG2Gj5fO47XlzR4PvsrrGOuO63OTsR5bwxbrvtlVne3JfBv0VXDC8z3Vbv Y7suunDQI4dt8oD/fl0nxilPnfSF6uW34scNVwL3zzXfT3bXFzV9dV0J59pCyS+Pu/Gx56brF0a6 tTfQeqnOHfbRZ+b6vsqLVx7f54d3NXCuSss+3FPrUNTzrzOFQlK7fM8lI9dN3hn/709WQXjdx3/f Fyv5HzkiRKxhncVl1JPNqVoiQLylil7164v39EfBuzVugMS7i/UKk0HtCYgdHeyfBwMXP4uxZC4B PBr3bFK6Co6mhIR7hnR0ho34/qEGfSuUlkuelkOOwSwFwAGZ2yhUIf8l7oXkMWHauJOklTEwiDw8 Yu0g56klHq9fv/uOA4FBxEZgToa6gWASP/QrLgIRit5yXGIY+KkEppF9GpTj3W4Ixzh+sIputKP6 tNiWNu4EgeTaI63IqJzoDZKP7tNhEWGCyDsukoprcSQiMZXFEE7SgNzSzSMJBTVMTmuEhXxKIq94 k+OIrF9wpN/t6IejA5Zyi9dTpJtyA0o8qnFWblxl1EYmylcyDZfDmmIoZ5VJC30Jf4cMJvsyNA1c cpKAeAzaS8bkDh/50pRKnKMHAalNaSJRmQtrSSu7xDA/gvFAafTmN7FpiuAo/sxyqikPR5rGrnfa Uof4lAUrsTeh9ySwnYphET5RuUqN9LB6b6wOLJHYzYOCzkf6DAlAAxqZKwUEQ1PrmSyzaM+N+q2S eIFVRK8pv/lV1ILGI+YRS4q3bRwyK+VaaVfSV42etQ+l/9ygJSP5v9VwZHMn3VI00/AEebrIjwbR 5ANh2AeroamfOXXhsZrY0/F9w6gbTBAQTpPVEBZQDUq1UxRtWtSnEqyX2MQqzZZ6MHbuEHlviRlX Z0EoDJYTrLzbTBDrUYprrrUGSE1rIbvK1uQYA4fu+YzTPIa6IjX2cqPQzGeW0wWS1uNwQGqpWhvi lKZFdVHn64yxeLO0zrLE/qCMlexoWavJv56Wef3ZrC9/Olk/9UgYvCPL8mx0ul5gYlXxiWLObFsW Nkn0CABp3RMHS8bftlZYiAWVUncLt9tQVLf5EEpueudVyt5VrMpKkVkNq8HZijKY0VVVb28KWpP6 Nq0vUa2RIKk4u7lXqlpU6UwfN1tPgrC7a2SdgPEV3y7dt6z8zS855zPElAb4Q1a1ITobjF+XbtZ6 Ft1ofRfM01HmUXfpbei3rElP/DjWdCKW2Bfp+2HVdditMWaq+AbqxZC2qcY/ZKqK30qP8+YYxDCm 3VWFNkunQhWn0nzni3GcQp6cF20PvqCQxaRGhuo4yDsET2Lf60o0TlRO/pXd0ks5JUJDhgyaHKZy KhsLUinnMpWaJeSYd+nTIQoXuTWD8qfqPOQ1dwyTPrxaLCm5QNESepCpk8snDX2tQAv6z8fc41Qh Xc5Jk6qZGCm0fnv8S3WStrloVKWMI22Sg+z5Na1C9YixPN4ZC4nHpQ7EqUO9T7Ut46e1Lpmu0yNr 4tDa1FbmNZZX3b5YB9vXxZHaRJ45bKsq2NUWXlCxk03qQSc3zXiNJ7yG2dROM4ja1a50t02ZyHJb u8VZhtKrYR1ucXPaz9nk800IzN5ziHHW8HM3bpz9biNvc59Ilph0meHMfDM5SPv2NzcRnr+oSm9M kWJ40Br+omMvfN20/ixzFwF843qT21LqBh+0wd1vjKf4jQCXc+aIyu5rg3njIM/hxZmt8JOnj67X ECfzZpdk6c7T3Dre860gZfObJwOj/EwTW1PeaqfTe94VF+r+pO4hox/dcDoX+UhZmGSXTzxdSmf1 gKveIZpjXW9c/jY4ffUxB1PaXWM/6Y83wquro31TbpdwkX/Wc3jCfUgo7nScFtrCopsc73dAbZTF TqSixTy5JOn7SqMCZJmX8ewfTzyEFn/Xxgf8wLl+s1mXO2UF89XFl+9i0Me4+YvIMK6fB70rDz3u cCFP9mUrvIe3gzknIz7xZ5ye1qtcTLKWfeu7xu7RyGofmW5H7g53/r1yKGNYok8TiscL8+jjZX3i +xWBGU8+W8Yf/enfFIUUNu3q68il3PPeEjhVO1wR5OaDl3/F5+e37waF0A0z0fxUb7oeKIDOrns8 Qf6cTP+gD4uUhPDWbqKkhcDMyRvkbcd07p7ERxB4zusicAFFxKEkcHtILnnoDPkCUO7OaGIkaOTA 7km8b78+UKS0bPhkgvA+Df5CUOOuTPXWCOJU4pZwEN2eraxgMJ3+SAZDbgeRCYM+aZP0yDMcafaw b1qAyX6k0PYg0InsStqMb9SScAuxY7jwDJTsifUszfCqbPsUqpYybQjJb/IgLJbYDPgiTW2EUAyb Ds4krQePCa5g/s4FNW9Dpqz/sEVkMi/4tgoJp0kOLSad+vCSmEnFkIvg7s/SICzPvgMMt20KU6Uw BEmn8C8PJeXi/u4NY1DPUMjjDmQTs7Dl2A8UI+zAeqX+UFAUKQ5mJPAHU68V/00LxfAEpW/qxK8F G1AD/yn7+Oz4EqwXVc4WOVEQAe/rZpAaSfAZR4kHJRHBmjEYz2okyoEPSM8bvUrN4FCvxpEcTe8Y McMzRKAFjowb8a+r4qqhnusLrqbY7DG2iCcm7u6LzvAd0/EaNxHN+mCx+ur0PCvKwsvLqC+7AmsO OnAd+S0Z6abz3DEey6+wFlJCfotiqurppPEet8b+hMg6PFJf/kAyI4mw1c6CIxMHJYsLH0PtudrR MkqxDuFNMKALEhORxgbDXEChdewvJKGRMKYR6RBE2XRhKHWSGwuSvMDrHN2rkaSSf25RGCcnE2PL +U5ytdJmglayG4cxVOgO6tLN1SyPNsxSLMcyGr2HLbEy3gxELWcjLtuSJYvyuEZmzqrRpl7LK7cR L5GSxQ7KEN0pDS0qPNzvxHIytbBxMMtxqMpI25DRF4uPMrssDC9jKwEtMiXzFVnR3kQzB/+P47ow K4Pt9z7TLXUJ+9hMDSfyCGOzEYvRL7IN0BxTyMypMwFprsgMKinSW0oJw3TT1oDTOGGsH7cSN+Vp anCLFtsw/ucEcynJ0B/tEOgeTYkmieocjTtBUzbvwQmvs9RqDAsxz4687d42rSH9Uj+0kzVfTqrI 8xTrMz6zhj7RIz9LEFDuM6D28zwAdC5d0T9JRkADc9poMjkLlCgWND8OtCflk0F5BkKX8j1rrUIn tNeSLUNb0z419PD8rUP1EhdB9IVG1C4ddDTR0kRXB0VT9DrC7UVblDr/bEaNkUVptG5udChu1N0Q UUctVERVNP8sMUgjh0d7lEhJtDuPlEKXVD+htDZz1EmtJUmVFEbvrUqRVEpjtEtxdP+29ElPLkOv 7krF9ENzqkzr8EzRlEDzC0Lvrk3dNDVtFEr9cU7pdEBR/upAgbRO9dTqvtRLa65BAdXuBBVBT+I6 89RQIxS9HJQ+GbVRB7LD8tNP33RS1xJRE7VQa2NTM9UpsVMcLhVTQfU2P1VIJU48UdVUHZW2HJNU jbJVoxTt8JRIJXVWUXM3czJWZTVXBxXrzPROWfVXwVM52XRYizVQ8c7mKhRXf/VZH6JZpTRaZ7Va FXU9T1VZS84nuQ0srhVUwdVTaZEziZVBATJUfU1GBVVcm/ELS/VYe09TtzWmfC9Nd5UBs5ReGyzn Iq5bBerXzPUzk85fmRVDP7Vdk1Bh2uVfYwo9u3JfQ86I3jIvncuYKNVXj5MHwzNiqQIiOyM45xMP mZTB/vxJBJu0Y4PFJRUSTJ0xiYaPBaPNE4v0XlMWeJZL2OCUl76vvWyTYl3WZq0CZ0O2olRILs+y 9Tz0F4P29lg2Z9W0AI9WakcwjHyWaTFmYwqKTzNLFTG2NB/lY82M+66247zMef4TCJ3WaJkx7Vor bZOWbD2wPZ+PswRDY9aWF5N2sd52bOO2MMeMb11VSvjmbmPWavem9JbNbxlRGVFPQrn0KLUSaIFR V/ZWM+F1cZfwBocxYSXHcgWxLqluYk0yc6nQk1b2H9PzRLl2GeXqO0NzZm8RPkv3EUXtPIsHb3nJ 09iTLNVFDxmJ0WhXY+tkYwH3ewiWo4w3NyeWGB3x/uuKV4+Ed5cksdl6VTWL6dWac4t+thPtKv32 8Dilt3Yx00WRUDRMTGh6c3jbLnCdl3HF19Yss3yvKApPdp3wApai0O1+T2AHExGt90EZczFtVYDd CX4fF3NL5Xz5ElYVpTgPWBzlgbOIxcDWk4LZAIKzjKsQ2EorcirpkWSP8GL5JzMCkWn1cR+9Fj/r aiQFF2XfsYRdWFmVy/0iEk6dyibhkSgwEi1GJIM5DHUPUobbzSDRsWIVg4d5coitlYZTck8RBqvA 8jJrlDDKa2r3VSgVkijbCYW5N1KZcjA+64eR0nJ9WGfJKzTCclS/Umx1eIwJE26LNr4gdY7fOCDh /kE5MVha9diOlzZdV5eK2TGQf/h/O9eAH+p6EbmPjfWKV7h5x5XsFjmErbFuNxPZLFmSlXcU+7dB 1feAvNKToTeT/9A5yZQ5PXmNrZOTaXQ8a7WVA9UJR1mE39WVb9fiXleWc1mXd5mXe9mXfxmYg1mY h5mYi9mYjxmZk1mZl5mZm9mZnxmao1map5maq9marxmbs1mbt5mbu9mbvxmcw1mcx5mcy9mczxmd 01md15md29md3xme41me5/k+x9Oe7xmf81mf95mf+9mf/xmgA1qgB5qgCxqVTdegE1qhF5qhG9qh HxqiI1qi7ZmeK9qiLxqjM1qjN5qjO9qjPxqkDUNapEeapEvapMO5AAAAOw== ------=_NextPart_000_0000_01C5B90C.B4B89930 Content-Type: image/gif Content-Transfer-Encoding: base64 Content-Location: http://www.uwasa.fi/stes/step96/step96/tikkala/fig3.gif R0lGODdhDwLGAoAAAAAAAP///ywAAAAADwLGAgAC/oyPqcvtD6OctNqLs968+w+G4kiW5omm6sq2 7gvH8kzX9o3n+s73/g8MCofEovGITCqXzKbzCY1Kp9Sq9YrNarfcrvcLDovH5LL5jE6r1waA+w2P uU9vXv02793ZfBn8v8cSODKIl1dTiJPYx4jylxDXsvgxOVMpebhz2chJuRi5srkh+kKaYlqa2bkq Uona8WoRqzJLUkuryqrLcQl6mhtySwcMI+xIvJtcsfnpC3n4iAAIneccEL1g3IYtTRy4Z70NcMCd HT49nqieW949fn1X2K5M756B/v6cPu2OXceNz1w+DPgGwjNILtM/fvoOznNYDt1BgRQTMmwYbdDD /nr0bhVE+JEaIIsACyoQFtJeRXEm+wXUJ5FkHHnWwLV0aXPgRo7KavlzVvIbu5qgIiKDAFQoQnE4 Zbp82i+qSpYwDUb6qZGoTm080/gElnWoQrFL11k9+sBsvmZbabYFqXQl06ksz64Fq+qcyKVd630t m9cs3bkn8dq9500kA6WA38oVPDgrvIaEIyd+F66vX7SLkcV9LFYu5crLRLEdPPU0abph7QXG7Hks 34WcNeuaNelz4ba7RbOu3TmzxeCoh68mnFt2b6Y5jYOePdK2bdyxeT9fPhr7RMQXi28/jvyyzJvf nVMzX917TOldqfMtr3o15OsayJf//b75ypC6/tEnRJ+fcgLtxB4r7jXQH37aOacdVzfFd19q4tVF nlv/bQMgggLuFl2BHJlWTXoMKpjdiCR6cBGEyd0FHYsUCGZTZSuqFw9wHvbBTI0BWtdgaAueGIyO xJmY4X42RjhRcwkCKeGRN6qRo4tDgrfkfD2+h5SIKoqI5HeowHgWfFx2ueSTBh7li5Vkbkgkkq0R pKVhjl055Zo8PjNZN3kW9+aVWJqJI1uHLbglXC2KhtKEWs0pFZWyGVrif6/5WCKEXQLaiFFh7TUo nZHW5ZR3EiTF6XhNUUXjW4w6uqmhpaYKE6bJ8Ddgd6yeeulHP5ZWoUSMvdRmqx2aClWoVQ17/mus st5mH05qjoiVA8CKOkGzviq3XrBkPdTOs+N5G6FwywZKIKh1olYUZxs5iGyxumVmKbHBQbprsfWq NW6++u7Lb7/+/gtwwAIPTHDBBh+McMIKL8xwww4/DHHEEk9MccUWX4xxxhpvzHHHHn8Mcsgij0xy ySafjHLKKq/McssuvwxzzDLPTHPNNt+Mc84678xzzz7/DHTQQg9NdNFGH4100kovzXTTTj8NddRS T0111VZfjXXWWm/Ndddefw122GKPTXbZZp+Ndtpqr812226/DXfccs9Nd912342hhninndzeaPft t9nw/hn41nNAV3jZ1ZgzYGGOXyhu4jOH/vj4t1tVDlHkksMMDebvwtUoV5uDfJ6eF5q+5+n3OTn6 yQrhmffjQ6GueuuTW4V6YrGrPrvtNQd2uu6p796777efFDziw6devPEx91667KDXvrzzLYMVu/DN N2+9y97sqf30u4/fPcuXSZl7p8uzXr7HY37O4uWwt3/9jsdOerj8tNOvMpqutgib/VWPfwQsoAEP iMAEKnCBDGygAx8IwQhKcIIUrKAFL4jBDGpwgxzsoAc/CMIQinCEJCyhCU+IwhSqcIUsbKELXwjD GMqQCKKbYcXaZUPX4TCHN+SPD38IxCAKcYhELKIRj4jEJCpxiT4kGhOfCMUoSnGKVKyi/hWZODTC UaiB7HuZjTTHvy7WbxkQFKP5tCiNMqJRZq9YY/vM2D83tiGCcERZHRl4R5PlUYF7HFkfE/jHkAXy gDUUGfvk+EZExpEgEhxkx7qoyO4V8mOQnKAjN1bJRkZSj5vcpPUumTExevKTozRkKTWZM1AiUJU9 ZCErJ/ZKQp7SfZ6cpe9iCTE42tJ2k8SYLi2IS4f9soLBZFgdd8lLZF7smBcspsKYCUxltvKFzkRY NQ3YS1iWMpuju2bBzLhDB3pzYLoMJx7HGTA5XXGd7EwiJ73YmHbKc55CLBk6/yW8LJWLEffEVD/7 ZZhR7ZOf0jQYNxuGF1n05aAE+6e+/vJZLc049EkTHVdAAVrQb2Z0YQn1V0U9xNCEdRSj70zZRT26 UXKm9GAjJaNEV5pOmGqUcReQqRY+WqCQGhQwjHxpSX8qwIhOx6YotSNPe+pTkuH0Ri116VCVulT2 0KQ+UiUqSe15VKQm1Y9Wver8VApVoFIPYFF9KlfF19Cu7qusC82qVs1qyrBirqYgVWu+2NoetL51 q3EVJPfklY0jmFMOThqswPC6GZpa7hw01JUi0OhYkdpVVthTFkmQNwRSTVax0toURzdLUe4BLo1C GC0itBibz5LOreATCiQyi7ugFgMt+iFfWlfL2WsEdqylxSxvZ5uWteR2ppQsy1eJ/gdaxBx3ov5D 32+rpVMvIJZZeq2sbYMAPNnO1n6vDSo9vwveHNEyt6/Tbe2SK9DsVJRLmDlu5sIL3ytmaby+Je15 Y4vd+g6QsJ3dX1M1ZFgz9AK9QzXuXOdK4AjQNsHXBY9tgZMtgs4Xk6y97IEZPGHzPne7/SWHbCFs q0BlOJQVphAOMRzcZtlgwdNy72Kj+wR1obgj1uVQOGec4oECV5/rimeLOyFjjX3PxhA2QoAtgVrN KUrHXiHcdDNVYdpAtggh1sSfDBvkw/XEyTjeRY3nwuTSwgax/nNIcPU5mS7/IMi+/HKMnpxeDOE1 tWnGEovlTGMewxkN/71LgwXr/uf9GqIzkuKuiaO35c4eeafDNR0zAP1nQ+hlH3ojTY14cr4wq9TA lRb0HI0caDvsENGODuieRT0vMLK0xPYVamPFXGVP543ObQWwqhl9YO0iCNSwHTWXKc2g8iZas6cu wzqkRVcqq5nIjZlw5zTMPBpzC8bSGfKuFUplUDdDwcDubvBmFWFdP5PVkV6MsgENUb2xo3LL1kOs Py2x/4rb3K9GN6FTDB/Fttvdmjbml/VbhX1fu17pszDxNkNti5Ib3lcQeGm43aZ8I9zhZ2r0hp1Q 7BF32k0Z5zfFgSzvizfh49BFs7NufSaS25pWMQbimZONix/uFt1NbPSPUyHz/t4qOueCUKK2keji UQ3DndAGwhJFjmSi91rptli0ibH77lrFaK+ecHrUH2t1fgyz6WG++oq9fnKunxLlOJ/lVTopdhTN WXROn3krHnGVOc7k5lg3+ySL8umZ5J3hZrby//ou5fiJ8u0ZkTLg897vXzxqcbqd+6NGQYh04bnx 8G5v33OAd/PGnfJFv/wogAfseEz+U0mfeTr2LlzLwlztWrZ8a+V8eg0nXKCVtXzsYw9tsnu76vv4 huYvP+beY961ob99mhv3ecbhPk8BDNBjVw54nZzrRSCYOkRaTWnXd3sYJt9t3PXO90shJUipz/3v l297pCveceg///JJ7+rX/moZcnJn7C/GP/DT68/5JqhNdstvfqHnYZ3nVDnWYfSWfsYXfrI2cBrH WYE2f/pHWq5HgG/ngI3Xfu2VfX/1IrNRZ2BWfwcIXW4kHGjiYbiHgiKoYB3YXGmkfRCof9pXblKn Z812bQn4ewvIgBzCbe8TOum3d7mHSHJUgtiTP5MXg4qmXN6HKs92ZbzyRfNggnKHeACYa6lGe7QG Zoa3ecF3cusEfzmogfWnhVkoY3ohCCwYhdvGhMAHhBtXgIpHT/MWJD0IRiGWcPHlYGInT64xf5zH cKoiiGYIYprSc5F0ZfnzD1/IbH3IfH+YgvYFd5cVXyGSgtI3gFsEWHYY/jnZQhuyV3neNiyO912z dn6BeIKZKGzhhXTg13tJhohpAYqA+IZBJ3WQ1R2LM4azSIWgCIDgNRzGR4HCVX6rSIRX1y6LuIu8 yHmhJoRq2Im2coS9iIm0uIgv5oimoSeRqIQ1ZWeYiIKi53acOILH1jkymIqHp4k1CIvd0m0KGIrx I4TuESWZlo7MeHvXuI7leIbHBj6Z6IuqCI5rBHbvFSY8dnxWaHEQd4DLyI10GIY8eG/Pho85SIvq F2C+9o4ASYwCOYFDuE+uMI3mh4ovKImKdIciyY0DKY7RtoPmwok2iGwn+JC+QX27RozfB5A7WQLq cpGhCCodiXi8UoGq/kcRxfiRLVleIPmHsohmvyiQ+riUMNWPxWiJqed4qMKCaqiCbpd9PGmLtDcl kViL4ucC/hCMkBh9r4h2F2hjUamVXqgjJDiJvPeLiuiLn3h9kiZNugcn3Tcv7ndnvDBIsyeWkTdK ysRNX4KXiulIhnmYVSdqBXV3yMRWgVRWmPmSgdeTjrlm7HKZnomYouSXZ5l4bVd2kHR2hzmSo6ma pYkJHgGbh/gXOkV2s0lYjAmZn6ebB7mCsCeHdLmb8YeNNadsvWlVRcQEypkEzBmWh7aH5FdPnwlA QbQERpSbxjlyLmcLtMly95ZqkuedKWF0w+lx2klWRDVjZXYu4PAh/ubpNNSGY2VIJC0JbripODY1 n/xXKcEobfjZaoajn6gGh8MzFkVJUKhZoEqgcmOAS11WhvIQkWxAd7IAn5KZSwNKoOTFadEJZWZJ jst5oV7WoNVnaNWjPuBmlyVqLA/TcZ6gguumfiIGCwpadyNacQ4aoYmYVzVqo1/3o/zyojAKnr81 pF/no0GadEr6UDKlZuxZn4yXaCYKoEvKomDwSk/6jY3onnlmojP4mWAKVo7gA1B6P7tHXXWIcVeq o6dQpmNCnDmKodu5TCu1bGZagyr6pS1Xp2T6pl25cjhab0S6pn1KB9S5oOF5pLGppoVqMYtKmDtX iZNqRY1Kp21G/gYsRqmbGkUWyKeY2qasJ6hiAKkvqVqZaktV+p5sOqOn2p3lGQyHxaqtOm79B3Wx GlOfKmQbJXB9NKsiqqtCZqu3Sgkx9aulKmAZ1au79KsM2qzIamzKClvFmp7Bikk9Oa0okp7H2qxn 0K1/qa349KzfGq2oSq5Yyq2CVK4klq6ks66g6qiU9KqwBgviCgXQ6hWEcG68gFLtqq6tsK+Q51WX +q+4mq0CK6TjylUAG7D3UFTxakgMO6gOm7D+GrEGe7AUu1YKK1ekuk0DC6wd67HTRLBK9a4RM6rJ p0cnm6H3eq5r4KuQpk0uq0PVhwTglLDWarLUymsaq3AWi1U8/tuw2PazNGtUQjuxXElZ+IqvEhau MguNL/t8UtC0H1qvN8upWau1W3tEK5NHLMq1YSu2Y8ukFFO1Q5d4cro3Z4u2bDtoheO2nRm3QAq3 Y1c1c1uz/Eo1eMtJH7u3Uuu1fjs1fCtWqxc1hBu0CHu4iAtVifm3m7N1iwu5gNtmlBu4lkuyksO4 Lrq5hXuTUNO5nhunTBO6WNWWoIu5Y0R1SpOy8Xm6rms8PiG5sRuLqOs8meS6qes9uotQvFs/vlur pEQ2pWtSKPk0rbu4Q3i8wLu7RNk0xLtIhPi8zKu6MUm9VXW9l0u228u9WxtG3Qu+4auHmpS2PNRD qmq+8Za+/uvLvu3rvu8Lv/Erv/NLv/Vrv/eLv/mrv/vLv/3rv/8LwAEswMWle6TmlhA5wDdVofp2 jAecwF1gIXFmqrIGvQ/Mhgw5wWz2wCObqOzWwC+3wWFQZh1lwMhylReGvCFMfhdIOciVixG2eCms wkmqcdFzoE3iXIoRcTN8nb42PzacVZKRPFfIwyIqjb5pwB7MwNdVwfu7HjecPd+YxKWTbkVMBWni jCWcIkEcfOFmxVcMxAVHLFMMgyr2xT3baS08xPLhjGvcxP67YGn5w20cxUs8wWd8nCLYmq/3YHTM xzKMx4SpPAa3xuujw7zjx29cvxFMyHUsIzmRw851x4E8/plVNlUuPHeL9VVeTMnN2WN2bCwtTBRv 2cnSlb2lPHyoLKuqvMraZrqnzDGAHJ5nJctAU8vMlliVfMt0I769XJDB6cvb+zvBTMwJVszdO8yE CMtLeFrO0qM0U4i77G5vu4lTykbRrMjgimRE9szX/HDd7Ac+Vr6h+jxEyFf8hYXj/AXZzMGja83h HKi0iqre/LDoTBzSbGX0DLKmuXMgSqPwZK/wjJMB5KUA3a8CbW6WWNDNG4eYtlkaHHjLjG8M3dDv ac96bMZ8JtHoargOfdFbSS0sG70VjXA79s0LKdLF29EenQpOCTtpQi7VS9K5jAn4J0D6A7MbDcG1 C1c9/mfT5/Wc5DzS1crPLk0+vpnTMi2uXeWTNx3U7azSt3WIP22gRBytOq3AxBUKqOU5Vp3SIjvT NI0L2tx57PzU9LXSLL3VzDyOSR3Vq7uqW91J/MkHZt1wtVRXa10fr8EJ+AzORFtt+jlliorVgNpX cG3RY/2bs6y2YA3YBbZ+i11VR2tNVEmQTGXXU5DZNJCll43Zebtq9yd0n524lS3akj3Zpe2z57ms x6hkQ1q2Jr2wiut3v4zQ7CiSzsrJsOrYjz18jxemdtgoXv2mdqFym62zvi1pRbqh3UjEGXfBSYtb tD21UXynSZY+MonaAi14JIfcJbvayx2Ek1zTzm2S/sSNprkZhNsnZr2t3M08gH6cynpceQvX1jvG eOzd3oet2UhN3nq9ccOoV1Rtz0B4at/dw1awlAgs2xyK02Kqg0UthhCOpPxNteDI4MUAwtiXogRu 0gbu3YXdclK74B7K2Vsa39CivPA94X5N4S17tTqQXS/e0nAIxfJc4bJH0GiM1tRNtx2Mec49XB2X 21Db41Q1mdNnBxNNKC8aL/tdXE8r44zM24EZKbBN5UNLYUj72ws834GajMbLl7tNrI+UquWpzqUg vcw92l2+qAjO4z6OdcuJkA7cgHMOsVueBeir4ZJq53d+Wi4O5GZrubF922Fn2B4ezny+zbFMuWSO /uZnp4hxnNbbleZAKuhQArgL8d+LHh2SzmWITZsfaMSMnrNc3uik/nOZt5eDPpOlB5y6bej4hLG/ 3elWmlBFvqb6veqZ3mS1Xt1ELoU7jtJofKS2LVnATs1O7o8Ieut4/uyBDudGJ7G2vuv2jeMR3uAj LuJGVu3VvesEd9b0Rs1RYOpr9e3ljnHFnuHpvegBN+1/uud4G4VInuNXLKxZPe+4KOfb/jdbQO8E 6bg/zjaY6+sn/edGneprU5hTru8SLOpy7TXWqa/jGVnbyXPjjunfWero6U/YabPATPFY65w0nprT SXNE9PFtpKp2epvn3n9M9hORfiSzfgwoZ/Ma/i2bDhLzOz9daQt3gkt44ATzdrnXB1/JNNzvgjx2 tQygaU6ZbAfIVVr0pmyZkUulK8xvWr/03EF4Sc71EyeURdfdw/3evImYYB/2EW+hUT/l7lmN9Y3F 8T4k8Ch9502G0P711t50Xd/2MW/t8BiA9e2C3R6Zdu+CVDGIgT50AK7klnr23ujy4t2M2Kd5+tjq s/KPQEmNvLPDhMruNcoduIjifsr2FOykfKmKlo/3Om74AMaLVimVcV9YtX8oyKeBV2kUkg8iVrKL ryj3Cthvg1WEANSE9hn6E2qLUsmRTpHF2a6jXZokapmVZlbzOD/KNHWEEkj2yY/oklpd3A/3/s1/ ikwZjWjIwM0nKQqvZ0pbFXCZJDSJl9avp44P/gI/9+SOgfvfhITNTgQQA00Tov0pt1B7F1e9efc/ gkJpDC1qklRobQ/ObWFqjO1bTfCd7/0f6NIFiSDjEfmKmSY6Z06IlE43SijDSlpoUyzqV5pdlbQR rA2cBomJv2EbHpe333OcGv9ZYpqnC1Sm5sbjboatL6Viy4zrz0IQzTBKso7EzOnkRXEvr1PDrqcS dJR0DlJE5ElTEoWnI5Ty9DMRNfUgMIh1UrTuj8UWdXZI8ZRz1phXdlP4uLh4TBd587fwLOP5uVR7 u1Coi9HXFxd2eqcqJxtUtLqcc48bOvcc/jkKHt3n1XX+0BtFXNleQIH1xt26dc0gtHQ01k1yGLAd voVdHkKciK4hjUgX4WQMFHGjIJAKH9FKGG9gypQAE2bKRNFSLiDYOBapCciOR3V0RnIZdFPmTJ33 3t0p82+JSqUCeX1b5WeLwqUVt/V0NcpqTqDA5GQ1tbWUVyaqYLaaejZsr2EtTdbwqvKtTbAnv87d yPNYnLg20VL96HIVwb6D1dktSJgoqaHmtC12Mzcaz6qGG1NGiRizXsv8MjvWt7OyYsj5tKbNfLj0 adVyPWNsPXXf472MJ+eVOLseg82pXj/dXdk2OdyriVtqHby4cY7Ikyu/yFyokuHmoP/9/m1xduzm 26mn4R7yy/ef3gtb6a2PvHjr4dW3Hx/G/a4k8aUZEc17+m349Enu5/+/v+uIM+Q88SIrjLSVtAMQ PAEZfBDCxKqKsBqMKLwQwww1RC2sDS3k0MMQRRxxQAfp2nAYE0wkkcUWXfzJnhURW+akAl+8Eccc BWtsxLUu0xHIIIWMBB4ZCUvRLSNxVHLIJhUrskdGYLLRSamqvHI7JlGUsgQtWawOyzD7ojJIH1X0 kkQwxVxTKTKBfEMtNllzU846+bLTLwnxfC+/Pf2ciRs0YVuHTicJ+RPRQAUUdKng+gxTj0QlvY/H NNeYFMZHMd2UmiddPJBTzkIdlboJ/km9stBTE001KlWbZNXVPWEtK9YcZ61VzltxFVLXXSFl1NcQ ew0WVWCJxVDTY/EcVllLmw0WVmOfbVPaaUWM1toos92V1Wq3LdLbbzXsVtwty3WV3HORDVfdCNNt F0Jm4f10M3bn7cjeewGUV1/1+O1X2HwBRuvfgccV2GBqE5604IVVa9hhCgtFOGIQK87VMoovjmlj WTPuuDmIQeZv4pFLNNnOklE+TeSV20v2RJcJ1ljmGSmjeeOWa/5O550p9RnLnoFObegqhS5aM6SN PlppeZo21DCcI2b6aczclNphqqs+MuqtLfKaVzKxXlhrsGfuyuzJ0i7TrrENLnvt/kbbjvtnul/U 1O2B4bZbQb6tzttvogM/e/AW90778MIVBRzoxBVX+3GFI3dWHsezZnxyp4VSM3NRO3f3uAQ//3H0 gz05HfXUVV+d9dbTK91012WfnfbabWcd9mJBzZ33IXfvHXgdfw+e+LthLh755JVfnvnmnX8e+uil n5766q2/Hvvstd+e++69/x788MUfn/zyzT8f/fTVX5/99t2XjMizPLJcYszbnx9Y/Iml/32LO5Ub RtCy3/1YIj+a7G+A7GtIz/QnwP7FJxkk4RhOJsgVsxAEObrQDR8smMGIYCJOpGsFdJjDjL+s54GS k9ChKvHBa+QFfwmKFEMwgR/8/pxpHzP8H2d06Bu1fOKHORRdCsFVkHIQI0lJqoUmLnhCJrLhiEGk oUaWuMRmwKlLxqDiE7k4DRhiMQPGIeJAQniIGpaBKC28xHjSKBIVWWmL5lEiF8VQRjjWcY7/iyJK 1DhGpmSRhyXxxxsJxcGiVPAwfdSTFyBxxgracZEpyo0bmzhIPxowMYXslG5ioskdRlIZWBRhHz2p SDzGb4KajGAgLXhJMq6xknK0pBwv1UBEHsqMeZwlH3VJjYyYUpWU9OE5hufKr1ywlIY84xE8t0Mk LeONoIxHMkknyUOaJZhGtM8wjVnEU2rTl48w5DU5BMIfmbKSnBwHOn9hS2w2/lKQmXwO57opGUi+ M5z5DCA52ynMXI4ynm2spjr5ac1p+rM3x6vnQReZBYI+1J3lHCcye+nMgHKMnYzcJz5xscoqivCT C+3KRV34oUU01JE7kqVJLQrQj+oTka2UaBwd2qVo3qOkIhWNC8/E0o+2w6Aq7ScOY5pRc27RqGLR ZRQv0dSeplKUOk0LPDX4wwBFtaBORWJRK/rML4JUqQNlRQ9xeUP+ga8nCzIjPWIqzSnW1KUrdFRF Q+pW5+yCHwtSq1SLdla+Ks6vfw1cYAXrN8IWNm4KRSz1FLtYxz4WspGV7GQpW1nLXhazmdXsZjnb Wc9+FrShFe1oSVta054W/rWpVe1qWdta174WtrGV7WxpW1vb3ha3udXtbnnbW9/+FrjBFe5wiVtc 4x4XuclV7nKZ21wBUmFp0HWuzfLwpk5Md1D0pGECK4ebYmJ3pHjT7r7Gax3wgmtijV1NeUNyXuC8 d0vXUS945xuYdTmovrtdDrPyO6g/utd/d0Qv6PoG4FvapmCHTQpcDBxHPZQVPBXq7yttyA7H7EXB 9Xymc4YoH5qI80EbfMt3t9lQ3x7Vw0DcIIeBgcS9+qumJbmiU2uh4hI3U7cEnaSNxenVHle3PtW5 VJCZ4WLSTMGsSQxqcHnc0SrwWKs/fvCQ1/NiKleZmE2eATOffFHl4Bi3/leWTpZzyUxu/geE5lQj DIws4w+D2ba/Q/BWrbmGHtJnnNv0Ioi7XFeNZNiVGH6Fm/cMZPKa9YZIfSEjPShi7qYwO1MG8XZ5 E4wVM8jRbUbSO/kQDTHz9jUJfnTMoCQcJoeaXxMu8H8fc2q5xOjGxdnysKaj6s3Wmr3CufN6Y42V Atk6swn9rn5yzbIOhzdpvUX1rjeX5gzPGmbD/gioa8JJCBM7yu7BpbQ5LF5qt7qfT5FNMuyX0rfq Wtx40e884dptJPNT27/EA4vhl9vzbNo1/smTv4YChgDha92Va2uANxpvgHdDx9Gxd7V/vZWRFBtf B6w3kbZ6pzDvFyy9/oL4uFfU7xECBdAEnIi3Ya1xYOsE3xO/7WKSlWr1blzhaDAxLNadDTqJeqcA lM3Bc5zWjI3av5tT+crnIzg/3RvpJ772MVcF8lo3GNw24zde9gt1iXSGqfmy0XiB7lm4fXpMbWN5 1zt7uBeHvS4PDzlpv85tCqe911bvxsOkxOupul3ua2/TwOVOuKGuac907Dsmw82mOpP9tzW0dbRx zujBSz3hwLna2bXi5sf/jfIRly/MS8X5y1c+v56PMHr1rlzRC3VxEPHz56lVerkh3riuhw3rZybd V7GH9ndXg3UNnXvNh07Ifs2KtPHu+8yPniH8bjSVju/75IPevhD8/mJ6ZX/a+RYf6y6Gr/PvUpuu Y5+N3N93zuvHavGL8Y9A/1f1P9v4EO/9/Ex0ixYH4VM8i8MnrsHH6i9PllaN4f9kwP6kL2YaSSbY j7Oc7SDmzyDqrCUeDRxopBfQz8iCIf4MSpBsKhxo4SWSzCrAbrvAJB/WIgJrhBhahQThzOow0Cl8 IoyeapiCb+ncLYZM6CBOqkZcEJsSAfYwiwXrji0c4cti8AOZLQSHb9ZwsAn+jwkeEAcRULMc6SXI 4AaX0AphaV+UcArzr+5SkFZyTwpb8PlqbANfigCtEA2b0AS38PSaC9gE7/5U6PzcD9PgL/7McPsw jUmgsOxo7Q3TV6/k7nAAoe++Uk8QB1EzehAO6+IQzSu8+PCuHlERR6v4wM9Amu/fGpFPOG5JygsT NdEDzYz3XgcUsU0UfQf3SlEVV5EVW9EVXxEWY1EWZ5EWa9EW46YAAAA7 ------=_NextPart_000_0000_01C5B90C.B4B89930 Content-Type: image/gif Content-Transfer-Encoding: base64 Content-Location: http://www.uwasa.fi/stes/step96/step96/tikkala/alpha.gif R0lGODdhCgAIAIAAAAAAAP///ywAAAAACgAIAAACEUyAdgbavVyEDlImVcTMdtsUADs= ------=_NextPart_000_0000_01C5B90C.B4B89930 Content-Type: image/gif Content-Transfer-Encoding: base64 Content-Location: http://www.uwasa.fi/stes/step96/step96/tikkala/gamma.gif R0lGODdhBwALAIAAAAAAAP///ywAAAAABwALAAACEISDB4mx3uBLs1UWL4w7twIAOw== ------=_NextPart_000_0000_01C5B90C.B4B89930 Content-Type: image/gif Content-Transfer-Encoding: base64 Content-Location: http://www.uwasa.fi/stes/step96/step96/tikkala/form1.gif R0lGODdhKQEYAIAAAAAAAP///ywAAAAAKQEYAAAC/oyPCAGbr6CctFrjrt4M5/t13KiF5Nk1aGiC 3tiiUkPX9qzEMpWpHLtj2IY+yKcYrCxoMuCPdExaYroeUzo9RSO8rLHrQWIfVhj3oCOnceh1kjpj uaXiUu4+ccPjifr4BbhxNGgH4/TXx0amhqj4c4gBJpnoGNgY2fZIiXZ3JYSXV5PZR+SnVGmZGtoT VyRm2uaJBKuHqgorxLpaNuppJGrpAyxspSd3yfnFSOrySeVAnGPMPBo7ZCuyDPoFzQP9DTq97co7 K159uUe97bjE9wmP/o5O67X4q82N+R7dC3JcjVywc7dKmQB2LyG7eGF0RVpz5RVBhvVOKcuXCmG3 /in9XsxhSJEXPHMc12nxdWbTQomuVG5U0VEVRpLYUtqUBxLhyC3OwFmbIzFcOUJKACJTt7LWzV4+ cc4kejEPtjq4AuLbiadqTqgCQzqF6XQMUpxasxFbghZQUJCZRLJ9EjVpIRERfdLEysjtWbTCrIW5 uq/RjZ8KrxW1+9AwQJT0VBryNdhl5HZ9cw3zK4oxRcehJs1DBuUj3DR3s3lbCDqsvRSBAVs8rVoh 57ip4f7ZKNVqs2S1ac9lbbrmUtmiWxv33RsoIr6Hzf4O3qT4cXtMdAqHHlf69Ne9V9x+pB3SGy1Q xu8Q3z096tRl1TfZ7j6+/PnY6dtf0f6+/v3nK/n7N6PdfwIOyBuBBkYQ4IEK2pfggt1V52CECjYo YYUWXhhEfhhuyGECBQAAOw== ------=_NextPart_000_0000_01C5B90C.B4B89930 Content-Type: image/gif Content-Transfer-Encoding: base64 Content-Location: http://www.uwasa.fi/stes/step96/step96/tikkala/form2.gif R0lGODdhVwESAIAAAAAAAP///ywAAAAAVwESAAAC/4yPBwGbD6Nks9q02lW7U29lTjcaJcmcIak2 6iV6L0gjjgZDL27X4D3DcHIrRXADRA6XylTzEzolZabj8mi1uZ4sHWb7yPpMv5I4O3KJJWtt702U mWu8XxmuQ08ndeQuDzjG11blVmHV1ydHU9d4RlhoxAgZ2SU5yIenpqkHxnQZJpiSaMi2SXaY+gXj CeV0eOoKpdFq1smjVxUr2eqq+OX5GBiJU5z2G/aXHJGo5rz8Gt07XMmMuQpLfCxc6nRqbMQNmmdL voeL3B2d3Q3+6siq6wyGnB5VCc8+Lj8v1ZL7rty9duf2rDOFDeE+UOjUUTvIBlq1UdXsLZQ18WKZ Y/8OJWYUotDhLorrLE7M17CkuI8EmXGsCFBjx3zeDJqEiBHnwX7Tao4z2UyVS4/yeNlcqRLf0Xgd Ww5LyVKdvXQop9C8BjKh1jvSBGKNimro1kDuwO6reulqSLNqGb40u3OGRbQBm47NWY9S2rc6pSL6 Ctdd24d068r0q08n1MFbf62Zow2mRlpZm94UC7UrIL5e1ga26jXvLIEpaYrm7FG00lKOMxwe6Ohb OCA7gD6rTOf2aGP/IAuyNTKZ7KIkgeH5sGn4KNrkjgsXMTJ5sNm6hDvnF3w5U2uiPHc/HHPMZT+Z BsLtm5Pod+/X766PyP09bvnp67dnRD/I4/nwOxdAoV+efwECKNR93+llRyj/HZifgAPa52BYBRIY n4GfUNgfhha+pxeCD3aHoBR0+OAhiCNqiGKKKq7IYosuvqhhAQA7 ------=_NextPart_000_0000_01C5B90C.B4B89930 Content-Type: image/gif Content-Transfer-Encoding: base64 Content-Location: http://www.uwasa.fi/stes/step96/step96/tikkala/form3.gif R0lGODdhPAESAIAAAAAAAP///ywAAAAAPAESAAAC/4yPBwGbD6Nks9q02lW7S6d1jmdk43VaIdKk KONWMUmz5fzEuovXEdjjlDbByWhlFHqAryExxRTdmk6foiUlIZW3bdWaXJ6KlBoZM15q0+YzZv3O sqIy9xWyw9vF8LusTYPk9bMnuDfVJwJVhJVjMyc4p9ICpfcDBrM1FkR5pFIXKWlEqaTRyDUIuUIW 0sljwob4QcqVg2Po13qUioZn+Xvl6qXrlznZG0aIaMi4mEm3ylssS7ipRwwTfIwMWSeaDY67LUyb Ku0Nnq6u/B1Mjow9zZ67O32+fi9uS98+L3+mT5u6e+u+yDsYKFo/YNR8TaLTkGG4etTyOUMXUVa8 gv/7uAmc1c4URIKD8kgkV9KEvWYeHYLcyLFlPIUIZbKsSTOgv4w8B1LcSFKHxJoooFnD2NMgPIgx +82kmLShxYUaR96k2s1jTqtFXWIlyI5mVHycRi27aFaUWJ0nr8bcinUhWB6PzlaE2qtSx70Jq35L WaapUrVMxUr6CVUfYLz+zCH+m+coILLl3HW59vBUrYS0ApcbFgqMqy+fX7Eh1XmXLtN1Z4ns3AUW CMvG0Oi1AjvtJUwgeQ9tvdN3VLe6vQYX/ntzcuTGgTMP/Hyw9DjRp0P/oGb3n+rYtW/nnsw580OB mpt/PoNVeUefwHen/t39e/nWx5uhX39+di34ve8N7w9ggAIOSGCBBnZQAAA7 ------=_NextPart_000_0000_01C5B90C.B4B89930 Content-Type: image/gif Content-Transfer-Encoding: base64 Content-Location: http://www.uwasa.fi/stes/step96/step96/tikkala/form4.gif R0lGODdhgQESAIAAAAAAAP///ywAAAAAgQESAAAC/4yPAZCq3KKcDdKLD1u5T+thYPg95OkgI2qy EZcsKwm7tJzVmqFT87+a2V48YUwCPOaWGmMH0kNKK6PqsHUtKqMhbtcD86oqJ24VR85OwcJz1Lm7 WGRw+W1dfCvVfFW9L/IH8hdD6EdE9WJ419WmmGjnc+YihuQIedgECFi5+eSkU0PnkymJFleIxzKK imV52qqZRkpbajpoqwjL8zjLq7X7IDrIMdepJRxc3MPqikNc5xUKRW3ZJPjTi3lNp1yNFbw9Jo5c 8hEU/T2crlpuuxxGDCwqGyY3rfvdAo86/Hu/LZ67d9jydEMjbREkgbIkcTuoL1cyiKkqOtwnD6BB iv+uGmJsiI/eDn8XB4KMuKyVvmNJPAqz5hLmHokmUSDMWHJmzI6+yDHzhpIdIosEy5Gcd81PN401 wQV16qaZNqL+EtIkl4KoRIjhUjoT2rNmJ1DoCoV8ylSn2IxHzyqNGOgqSbfcdmL1yNCqXa3/5Cqs yLCp3Kl7+/oUGpjn3ZZUVxo79MbrJ7/15EWGe7fox6nHFDPm2wgZvrTjQI/NKbiyO5aXTLN9vHrD nDGn4IzGlFgyrpeGDe3eHBAnZLWKi08G7vRREMKpszJnLRqnmQ1hPb/Gqw52iqVp9SYHLnKijdzc ZzUDGkiqYZvcjcVDh039X18H77FyTw1+jvK9DSZve+ufA7CEY5Zs8j1T20327NIVV6/Md5onnbFX CyfzgUYJKJ58tWGHHn5YRoWceGjEhUxcYWI7sfRhBYguvgjjiuvxkeITQ21SI3Hj3ThiaTH+CCSO MeZo44tE6kgJjC0GyWSTTj4JZZRSTkmlkwUoADs= ------=_NextPart_000_0000_01C5B90C.B4B89930 Content-Type: image/gif Content-Transfer-Encoding: base64 Content-Location: http://www.uwasa.fi/stes/step96/step96/tikkala/form5.gif R0lGODdhZQESAIAAAAAAAP///ywAAAAAZQESAAAC/4yPAZCq3KKcDdKLD1u5T+uFyAaKYWk+6Ycu qMmtqfsm8Wjc18u3sgfRfSQ9W6eWCSKNlYjQ8XtGS8qfygqsUnwiKdXlxG67te9zqVmhnTwbuCmO V9ZyWTt9orvDc3Z9pKdR5re3MxMoyFe4+Ccm1ajmBRJDM5a4U4nTp2j3lgOI6YnHODq0JTq5RkNF CKci+nCjA5aqqpQJCot7lcNhAbtJlHr1KPtI+hmMvLq6Sfl7fOkKqszprBWbdIfcG4v9nJ3MS3k6 PAfeTSx7SX6OMUuSrhzNzpxZjHgtz8siaI++7J8QdPC09RK4j5qxfb5+aVrYT9oigOpKxUsYqKA/ Yf/UMFlLKLETNn4cP5IEySQUQzpeWhWCWNCYmV3zykBbKQ8VM3ce653E+W5bx4D//FCk9xMKN3po GgLtmXKowpv8nPqjajCktKMj7+mJFnPkUKQ/i7iEoUXjKSJnUWoqyTVrVK3pYA47QxFqUrtT3ZHd WLOiS7Itq9HNUtfcWpOSDEWMG5SxzZxdLyqF4mmJ2omVc+LIfIRT2skSG7My7PbI6I6mp9Fluhgy Sl9z3QLkO87hQaaKOfukjKfZFMo0R+ECBuwcciw7lXptQVh4qUj3rgIGZKv4LRIyt3/SxbshzUrH M7/hjt1Rcr0m1UNKKu7PX7SOG81/jz+//viu6+Q562IKJP+dYImA+x2IYILTLSjHgAD21+B+g7zn oIIWXviOghU+eOCGBCboIYYijkhiiSaeiGKKJhYAADs= ------=_NextPart_000_0000_01C5B90C.B4B89930--