Click for next page ( 47


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 46
Deep Basing Concept (Horizontal Egress) JAMES A. WOOSTER Boeing Aerospace Company Seattle, Washington SUMMARY: Recent developments and observed trends in the intercontinental nuclear threat against potential strategic targets in the United States have caused much in- terest in new concepts for survivable basing of this nation's own high-value military systems. President Reagan's October 2, l98l, announcement initiated the current in- vestigation of deep basing as one of three possible long-term basing modes for the MX intercontinental ballistic missile (ICBM). The same threat trends (primarily for in- creasing accuracy, increasing numbers of large delivery vehicles, and increasing num- bers of individual nuclear weapons) that engendered today's ICBM rebasing activity also affect our approach to the design of deep basing systems. Most importantly, we should avoid (a) any dependence on fixed surface elements to perform mission-critical functions and (b) the temptation to concentrate a great deal of target "value" in one or a few deep underground cavities. In other words, we are driven to underground sys- tem concepts that can operate relatively independent of surface support after attack, and that spatially distribute target value to make a nuclear attack on the system as unrewarding as possible to the attacker. For a combination of reasons, both techni- cal and nontechnical, we also should avoid dependence on deception of the potential attacker regarding the exact underground locations of critical fixed system assets. Recognizing the foregoing constraints, Boeing engineers in recent years have studied a series of concepts for deep basing of an ICBM force. Their efforts led to the description, in some detail, of a particular example of a deep basing system con- cept, and to some parametric investigations of the anticipated cost and survivability of such a system. The example is an interconnected network of horizontal tunnels, excavated deep under a mesa or mountain ridge composed primarily of unsaturated por- ous rock. Access tunnels are horizontal, but passageways for postattack egress may involve slopes anywhere between horizontal and vertical. Provisions are made for critical subsystem equipment, personnel, and materials to be distributed among many separate locations within the tunnel network. Preliminary evaluation of this type of system concept indicates that satisfactory nuclear survivability probably is achievable at depths that appear to provide a reasonable number of candidate sites in the United States. Those of you who have an agenda will notice that this slot on the agenda is entitled "Horizontal Egress Systems." I want to take the liberty now of expanding that title a little bit. I will try to speak in basically two categories. First I'll identify some of the overall system archi- tectural design options that we believe are available to designers of 46

OCR for page 46
47 potential deep based systems. Then I would like to go into a specific example that we have worked out in more detail than the others and that gives a reasonable outline, I believe, of two things. One is a system with near horizontal egress capability, and the other is some of the crit- ical subsystems that must be involved in just about any deep based system (particularly a manned system) that might be considered. Just before we start, I would like to recall several things that speakers have said today. First of all, Colonel Berry's presentation on the evolution of the threat is pertinent because it tells us that as the attacker's accuracy increases we cannot count on anything on the surface to survive an attack. As the attacker's ability to deliver very large amounts of yield to the target increases, we also are driven, as Dr. Se- vin pointed out, away from concepts that tend to concentrate high-value assets in one place, regardless of their depth. Thus we are driven to distributed systems by those two trends in the threat. Finally, some of the material that Dr. Merritt just presented in the way of weapon effects testing, and particularly vulnerability testing, of various cavity lining systems and that sort of thing gives us some basis for constructing analytical models of deep basing system vulnerability in the gross structural sense. What I will show you, I believe, will illustrate how that type of in- formation can be used in at least the very preliminary stages of a system design. I would also point out that in an attempt to meet my time sched- ule I am going to be walking through these charts at a speed of about one per minute. I won't be able to elaborate on everything, but I want to give you some ideas as to approaches that might be possible. An obvious one, I suppose, with which to start (and with which, in- deed, we did start) is the idea of distributing a system in the sense of providing many self-sufficient deep shafts, as illustrated in Figure l. In other words, just very deep silos that are capable of protecting the missile from attack and provide the capability, also, for self-sufficient digout and all the other functions that must be provided to support the missile. In the plan view, because of nuclear attack considerations, you are driven to some sort of a hexagonally packed layout to avoid the at- tacker's being able to get "bonus kills." In the elevation view you must consider some possible nonideal geologic conditions. At this stage of the game, about as far as we went in that direction was to consider two- layer systems. At the time when this was being done there was a temptation to bring in the idea of deception—in other words, having more deep shafts than you actually have missiles. Recent events in this type of business, I think, would convince us all that deception is not a very viable approach to de- sign. Deceptive basing schemes have a lot of ugly aspects as regards pub- lic acceptability and cost, but nevertheless we tried various approaches. You can try to shuffle missiles on the surface, as indicated in Fig- ure 2. You could try to do it in a shallow tunnel, as shown in Figure 3, to help conceal your activities in operating the system. Or, as illus- trated in Figure 4, you could provide a shallow tunnel for missile shuf- fling and a deeper tunnel for less mechanically intensive operations, such as minor maintenance.

OCR for page 46
48 We eventually came to the conclusion that, at least conceptually, this sort of a design approach would lead us to a deep tunnel system where the interconnecting tunnels are as hard as the cavities that pro- tect the high-value assets. (Such a concept is shown in Figure 5.) Al- though we can't be certain this is the ideal solution, at least it is a very attractive one. I want to be careful I don't say anything is an ideal solution, because I think before I am through here you will see that the work we have done barely scratches the surface as to the pre- liminary system design activity that we are facing in the next two years. Returning to the question of postattack egress, we believe the un- certainties (primarily in the nature of crater-related damage to such a site) tend to drive you to consider the idea of creating entirely new egress paths after attack and avoiding the immediate locale that was at- tacked. Such an approach seems superior to attempting to dig through or otherwise pass through an environment of very disrupted material the me- chanical nature of which, or even the extent of which, you are unable to predict prior to the actual attack. This crater-rupture environment un- certainty makes design of a digout system (particularly an automated one) an almost intractable engineering problem. Figure 6 illustrates the fore- going issues. There are also reasons for looking at the variation on that theme of providing prestarted exit pathways that are not completed all the way to the surface. This approach to system design has two attractions. One is that if any of the exits are not attacked or happen to survive for what- ever reason, they provide you potentially with a much shorter egress time, and that is militarily a very attractive feature. Referring to Figure 7, another attractive thing about pre-established exits is that we believe they complicate the targeting problem for the at- tacker. If he sees a system like this or believes that the system is de- signed like this (actually he would know it) he then must make a choice. He must decide either to target the high-value assets at great depth, which are well protected and thus require large amounts of his deliverable yield, or he can decide to use that yield in another way and attack your prestarted egress pathways. An attack against prestarted egress passage- ways requires him to use up a significant fraction of the yield that he might allocate against the entire system and thus leave large numbers or a large percentage of the higher value assets, such as missiles, power plants, crews, etc., undamaged at the greater depth. If so, they would still remain a long-term threat to him, although perhaps not as immediate a threat. Although this is probably an obvious point, I want to mention brief- ly in passing that the basing of digout capability in such a system also clearly provides the capability (at least conceptually) of repairing some of the damage sustained during an attack. This idea is illustrated in Figure 8. It gives the system flexibility, particularly, we believe, if it is a manned system. It would have flexibility that an unmanned system without this capability could not exhibit. I will apologize for the cartoonish nature of Figure 9, but if you think about the problem of creating new egress pathways after attack and avoiding damaged areas, we believe you will be drawn to the idea that

OCR for page 46
49 perhaps vertical egress paths may not have all the attractiveness that steeply or even shallowly angled egress paths might have. In particular, if you are capable of finding a site that allows some surface relief, the task of tunneling or otherwise excavating an egress path, we believe, might be accomplished with less of a demand on techno- logical development. In other words, it takes more advantage of existing underground excavation technology. So, with that as background on the preliminary search for a system to examine in greater detail, I will proceed now to the system that we have recently given some attention to. Please understand, however, that our elaboration of this concept was done only as an example. The motiva- tion for doing so was (a) to convince ourselves that we understand all the parts of such a system and how they must play together and (b) to provide a framework for the planning of required research, particularly research in nuclear hardness and survivability as they relate to mission-critical subsystems. Now, I have a series of four or five illustrative charts here which tend to start, as you see in Figure l0, with an external view of an ideal- ized site. The surface relief shown here probably is physically unrealiz- able, but the point is that a base such as you will see described would have very little observable signature on the surface. You would see a system of access roads and a system of access tunnel portals from the out- side leading into whatever escarpment was used as a host for the system. As we see in Figure ll, if it were possible to cut away and see what is inside that ridge or mesa or mountain, whatever you want to call it, we propose that a tunnel system be excavated in that escarpment which ba- sically consists of, first of all, a peripheral tunnel that essentially follows the lay of the land, the outside periphery of the ridge. For rea- sons which I probably don't have time to go into in detail, we believe that you would, also, be driven to have essentially enough additional un- derground space to provide a redundant tunnel (shown there in a zigzag shape) that connects with the basic peripheral tunnel. We have shown the idea of providing prestarted exit tunnels sloping up nominally at a 20-percent slope. This concept for prestarted exits is one that we chose rather arbitrarily for purposes of this exercise. Figure ll shows that for some distance (which would have to be determined by our estimates of cratering weapon effects, etc.) that exit tunnel would not be completed all the way to the surface. The access tunnels are shown again here in Figure ll. They would have to be provided at intervals, probably something on the order of every l0 miles around this system, for two reasons. One is that, as you people are more aware than I, access tunnels would be necessary for construction purposes during the deployment of the system, and finally, of course, it is required to have some way to get crews, equipment, etc., in and out of the system during peacetime operation. In Figure l2 we have made an attempt to show a closer view of what is in that internal tunnel arrangement. This is a view that at least tries to give a conceptual idea of what that system would look like from a closer vantage point and what sort of equipment is required in it dur- ing the period of postattack egress operation.

OCR for page 46
50 We have shown that the machine to be used would be a tunnel boring machine (TBM). I don't think that we need to select a particular device at this time, and we realize that a standard TBM is probably not a good choice, given the uncertainty about the type of material you will be go- ing through. However, some sort of excavating capability has to be pro- vided along with some means of handling the spoils, a place to dispose of them within the system, and of course all the other critical items. They include accommodations for the crew (in other words, the life sup- port systems), shops for maintenance of the equipment being used, trans- porter-launchers which house and protect the environment of the missile while it is being stored and, also, serve the function of bringing it to the surface for launch after digout has been accomplished, plus some sort of survivable electric power system. We chose to look at large-capacity hydrocarbon fuel cells as a candi- date for that system and, of course, recognized the need for a very im- portant and very elaborate environmental control system which provides ventilation and disposal of waste heat in this system. Before I proceed to a short discussion of some critical subsystems besides the ones that you see there, I want to make a few points about assumptions that were made. First of all, we made assumptions which I believe are critical. One is that nothing in the layout of this system will be unknown to the at- tacker, that he will have perfect pre-attack knowledge of the location of everything underground. It is very difficult, we believe, to con- vince ourselves that we could keep that information secret or even sig- nificantly uncertain. Secondly, we also believe that by the time the full range of threats against such a system has been considered it will be a requirement that the system have maximum autonomy in the sense of not requiring exchange of air or coolant fluids with the external surface environment after at- tack. In other words, it truly must be a self-contained, buttoned-up, sealed operation after attack. For that reason, as many of you are well aware, the problem of thermal efficiency of all the equipment involved, and particularly energy storage and energy conversion systems, is ex- tremely important. The problem of disposing of waste heat in a fully sealed system that is housed in rock (particularly the types of rock that we believe would be attractive, which tends to be a rather good insulator and not the most ideal medium for disposing of waste heat) will have to be given significant attention in the design of deep underground surviv- able basing systems. The previous figures have shown you a few of the critical subsystems, and I particularly want to point out that, as we are all aware, the post- attack egress issue is the first one that people will ask about in con- sidering this type of concept. It is my belief that perhaps the second question that will be asked has to do with the survival of communications. If this system is to have any utility to the nation, some means of com- municating with it after attack must be provided. As you noticed in the previous pictures, no designs were chosen for that particular subsystem. The reason is that we don't believe it has been worked out, or that the technology in general has developed to the point at which a particular

OCR for page 46
5l design can be chosen. Thus, although there are a number of promising tech- nologies from military applications, oil exploration programs, mining safei- ty research, and that sort of thing, that lead us to believe that some sort of through-the-earth communications link would serve well as a last-ditch survivable link to outside authority, we did not believe that we could iden- tify a particular design as being even a reasonable candidate for an example. Now, clearly you must have some criteria on which to judge the bene- fits of such a system of deep based strategic missiles. Our aerospace dis- cipline in this area tends to concentrate on these two criteria: how well does it survive an attack of the type that we think might be mounted against it, and how much is it going to cost? At that point, particularly with the last word, "cost,1' ringing in your ears, I would like to caution everyone that the amounts of resources that have been devoted thus far to this type of conceptual system design are by no means adequate to provide a lot of confidence in cost estimates. I think that before we are through here you will understand that all the confidence that we place in the estimated cost numbers that we have come up with for this one particular example is to convince us that it is not an order of magnitude cheaper than alternative basing schemes, and neither is it an order of magnitude more expensive. It is in the same ballpark. I mentioned the capability of creating survivability models, given some rather arbitrarily chosen parameters at the outset about the type of site that you will be in. In Figure l3 we show at least one model of the survivability of tunnels of about the size that we are talking about in dry, soft (that is, unsaturated and porous) rock. You can see that an analytical model (which is a lot fuzzier than that nice crisp line of Figure l3 would tend to make you believe) can be created which, for example, shows that, from the facility designer's view- point, l00 megatons of attacking yield are required to irreparably damage a single point on a tunnel at a depth of 3000 feet in the type of material we are talking about. In other words, trying to provide a system that we believe with great confidence could survive a given attack, we would say at that depth one point on the tunnel would require something like l00 megatons of attacking yield in a surface burst to create a severe enough destructive environment to render that point or a few tens of feet along the tunnel inoperable. As Dr. Sevin mentioned in his presentation, the tunnel system design- er's viewpoint is not the only one that counts. The attacker's viewpoint, also, has to be taken into account. The attack planner has a lot of un- certainty about every step in the process of predicting how much damaging environment he can produce at this system's depth. Even conservatively speaking, we believe that when those uncertainties are folded in (as you will see) there is probably something like a factor of eight between the two points of view. If, for example, the designer felt he had a system that was reasonably survivable against l00 megatons detonated on a par- ticular surface aim point, the attacker (at least if he uses targeting philosophies that we believe he would) would be convinced that he had to put 800 megatons on that aim point to ensure a high confidence in destroy- ing the deep tunnel target location. Such a calculation of target hard- ness would, in our opinion, tend to make any potential attacker look very hard at other ways of neutralizing that target.

OCR for page 46
52 Now, given a model like that and some knowledge of other nuclear weapon effects, you can come to some rough conclusions about some of the parametric variations of such system designs as a function of depth for a given set of other postulated constraints. Some sort of a threat esti- mate must be obtained. Some requirement must be specified for how many missiles out of the original number deployed must survive. Also necessary are some specification of the type of site and an agreement as to what the proper kill mechanism is; that is, that combination of weapon-induced en- vironments which will render the system inoperable. We have done a little bit of this kind of thing, and in Figure l4 we can see a couple of curves that are important for the type of system I have just described. As you can see, as you go deeper, you require fewer missiles to be deployed. Also as you go deeper, fewer miles of tunnel have to be constructed to interconnect those deployed missiles. As you can see, these curves do not have a definite optimum point. However, they tend to tell us that if we are interested in fielding a sys- tem that looks reasonable in terms of number of missiles and if all of our other assumptions about the threat, the nuclear weapon effects, and the survivability of tunnels are correct, then you want to be somewhere down in the neighborhood of at least 2000 feet deep, probably closer to 3000 feet. Again, I don't want to give the impression that there is a lot of confidence in these exact numbers. We did this type of analysis primar- ily to show trends, to see if there were any obvious optimal depth points, and exactly what were the trends of system requirements as you go deeper. Doing that and having some idea of how much it costs to dig tunnels and shafts, provide various pieces of equipment, etc., you can make an esti- mate of how cost varies as a function of depth. Here we are going to get into some Defense Department cost terms. We see a few of these terms in Figure l5. Life cycle cost, for ex- ample, is the total cost of doing research, developing the system, de- ploying the system, and operating the system for a given period of years. Research and development cost, acquisition cost, and operating and sup- port costs are depicted individually in Figure l5. Acquisition costs are just the costs of actually producing and in- stalling all the necessary equipment, plus providing the necessary base facilities, including underground cavities. Out of acquisition costs, just for curiosity, we display how much of that in our estimate was occu- pied by the cost of excavating tunnels and other cavities. As you can see, depending on the system depth, it is a relatively small fraction of the total. Figure l6 is a display of the same data for a particular system depth in pie charts. Please keep in mind that in de- veloping this estimate we employed techniques good within plus a factor of two and minus considerably less than that. We can see, however, the division between research and development, acquisition, and operations costs. Keep in mind, also, that in the research and development cost cat- egory we charged the development of the missile itself against this system. In acquisition we also charged the acquistion of the missile against this system, in developing a number which comes out into the few tens of bil- lions.

OCR for page 46
53 What we are probably more concerned about now, having been charged by President Reagan with comparing three competing long-term MX basing options, is just the cost of the basing system itself. To do that you have to take out the cost of missiles in both acquisition and R&D. How- ever, the remainder of this acquisition pie (which is about two-thirds of it) is split about equally between equipment and other items and ex- cavation. These costs must be charged against the cost of the basing system. Again, if you look at system acquisition costs for basing only, then tunnels, at least in this particular example, loom as a larger frac- tion of the total (about half the total basing cost). Again, I want to offer the caution that this exercise was not done with the intent or the claim that it produced a system that we could go out and build tomorrow, or a system that we could even stand up and say today is the optimal system. We clearly cannot say that. We have not done sufficient research to identify an optimal system. We have not ex- ercised all the possible deep basing options in this way. However, we feel that this example was useful, at least as a starter, in portraying to the community the type of considerations that have to be included in an R&D program such as we are facing right now. With that, I will close with a couple of minutes to spare. SPEAKER: It went a little fast, but what cost per linear foot of tunnel are you talking about in those estimates? MR. WOOSTER: The estimate, which was done in l978 dollars, I think came out to something like $l,800 a foot, at the most. SPEAKER: What size tunnels were they? MR. WOOSTER: They varied. Different parts of the system had different tunnel diameters, but the access tunnels were of about l8-foot diameter, and most of the rest of the tunnels we estimated would be l5 feet in diameter. SPEAKER: You passed over the shell game of the old silos very quickly. What disadvantage did you see in those? MR. WOOSTER: I will cite two primary disadvantages which we feel were a great hindrance to the MX surface shelter deceptive system. The first one is that producing redundant shelters in which to house missiles, par- ticularly in this deep based example, would be extremely expensive. While it might enhance survivability, the costs quickly get out of hand, and just from an economic standpoint we feel it would hinder feasibility of the idea. The second one is that in this country, with the society as we have it set up, maintaining deception in any system like that (with the possible

OCR for page 46
54 exception of one where everything is at great depth and fully concealed) is very difficult. I believe we would be unable to assure ourselves that we could maintain deception and be confident that we were indeed still creating enough uncertainty in the eyes of the attacker. So, since both of those things were looked upon as close to being "show-stoppers" on the previous concept, it would be nice to avoid them here. SPEAKER: It might save us a little money, but what are the Russians doing? What are they going to do to base their missiles? MR. WOOSTER: I am not even briefed into that activity, but we ought to be concerned with two aspects of what they will do. One of them, as you said, is looked at from the defensive point of view. Would they mirror- image a development like this if we started it? That is an interesting question. Perhaps it would be a good thing. The second question is, "How would they perhaps modify the forces they have in order to attack a system like this effectively?" I think in that area lies one of the primary advantages of deep basing. I say so because I believe that the nuclear survivability of a properly designed deep underground system will not be sensitive to changes in the enemy's threat, or even to some very substantial changes in his threat. SPEAKER: Your presentation was based on a prototype site. Would the system have multiple sites? MR. WOOSTER: I don't think that issue has been even addressed yet. There are some considerations that I think would drive you to wanting to have multiple sites, among them threats in the non-nuclear category. SPEAKER: Which means you have not eliminated the possibility of the silos; you have not totally eliminated anything that you started in the beginning. You are still going to have another look? MR. WOOSTER: That is right, if we are permitted to look. SPEAKER: So you have not done our job? MR. WOOSTER: That is right. I certainly don't mean to imply that any options have been foreclosed. We have some reasons for believing that some of the l5- to 20-year-old approaches no longer are viable because of recent developments in the threat, but there is still quite a wide spectrum of design approaches that we believe are still valid for in- vestigation against today's and tomorrow's threat. SPEAKER: Does your scheme depend on these 5 percent and 20 percent tunnel slopes, which are pretty tough to build? MR. WOOSTER: No, it does not. I think there are two main penalties for going to shallower slopes. One of them is that shallower slopes make acceptable sites harder to find. So, site availability from a topographic

OCR for page 46
55 point of view is much enhanced if you can go out at a steeper angle. Per- haps you don't even need any surface relief, if you can convince yourself you can dig out at a sufficiently steep angle. SPEAKER: Conversely, talking about a postattack excavation, I would think you might want to keep it simple, and you might want to make it horizontal, and that would limit your geographical sites. MR. WOOSTER: Yes, it would. SPEAKER: What digout times are being considered? How long do you have to get out? MR. WOOSTER: The answer, as far as I know, is no. SPEAKER: The answer is no. There is no constraint, but obviously faster is better. MR. WOOSTER: That is right. Through the several years that our organi- zation has been looking at this particular problem there has been a dis- tinct paucity of specific requirements. We have had to postulate what that system might be required to do, and it appears that in this program there is going to be a deliberate approach which says that we want to see what is possible before we start laying any specific requirements on the system. There is a wide variety of opinion as to what is desirable. There are some people in the Air Force who believe that if such a system cannot launch instantaneously it has no credibility as a deterrent. There is another variety of opinion, perhaps in the "strategic thinker" category, of people who say that as long as you can create enough uncertainty in the attacker's mind that he cannot actually destroy the missiles, even if they cannot dig out at all, he still has to consider them in his cal- culations of threat against himself. That point of view is not terribly appealing to me either. It does not constitute a very credible threat against an attacker who otherwise could bottle you up. SPEAKER: What is to prevent the observations of rubble after the attack and the immediate "zap" when you break through before you have the time to get your missile up? It is obviously going to be a long period of time. MR. WOOSTER: Yes. That question has not been addressed, and it will be a key part of the R&D program, I would predict. However, I think that one of the answers that is going to come out is that even a system such as we are talking about, which has a high degree of self-sufficiency postattack, is not entirely independent of outside help. For example, one common thing that is said is that we must retain the capability to deny enemy occupation of the site. Working that problem even further,

OCR for page 46
56 then, perhaps denial of surveillance or at least interference with post- attack surveillance on the part of the adversary may also be a require- ment for systems like this. DR. LINGER: That is one reason why there is a task force whose objective is egress. Egress is the big problem. SPEAKER: I thought maybe they solved it already. DR. LINGER: No, as a matter of fact, I think there might be some words that will come out here that would help. Are there any other questions for Jim? Super. Well, thank you, Jim, for an excellent presentation.

OCR for page 46
62 M t£_ o CO i CO v. •j < £ CD O UJ >4_i Z ^r^ IU j-j ^^ ^^B t CM ^* ^T 5 Q 3 i R • o £ IU CD -J .2* 0 DC ^ Q °- 12 K * TJ || i1 3 0 X UJ o cc I M H £i! Q. 0) 0) •• 1 S I b < § -i Z. £ M . 0 - 3 = > 1 5fc + 2 § | 5 z 5 o ~. O v Z Q i u 5 £ •u < 3 Q (- 0 CD ui 1- IE Z -i < X O 1 i £ s i§ co £ O O

OCR for page 46
s co 1S3 O m M C 5 ' ?s3 > 0) —

OCR for page 46
Tl M tO O ? in m G1 1 w "o C \ O 73 • ''/$ m / • »j - s /t& V 8 ? ' X/?V^7 o o //Q-Vr m lESTRUCTION If OR. THROUGH 3R.COMBINEC SCO m (A 7 T< * 9 9 > »~ 3s ^,\-}\ m g K l^w-Y? > r S_ > OS i\\\^ pi%w w Z ^ O ___ 41 ^IX^ o ss 0 3v|iH> / 19 CD CD a 5 -£H \ M ^ 55 > F S -"£> ^, > 55 -n > H > n H — i ^ 5 § c ( S M X 1 5 J trt -O M IXH >m 3) -ni,m ^5«? H C 7 3 = 2 • c O 8 » H' a X . JL ^ m3 O- , m c •M 2 ' X z ^/t ^ > m o CD CO* 0. O a l-^^ '1^ M / . — £* klji> 3D m^m O m ' W 9 r ni i-f > I !\m C ^ i 'i r-h O 33 V \» 1 \\ a ra m V> * VI 1 I ^ N > | Z H \ ^>*> ' , ^ ^ H ^ s ^N z / D s Z 1 -*M 0 m JO 3 Z m s ] CD f* 0) r a I 0 M ^

OCR for page 46
65 3 C O O O) LLJ .* U (0 C < U c/1 u < 52 - S o o ^5 H UJ UJ Q 0 a: 0 S o: oc a. iu 2 ® i' z UJ'H . o z > 4: ^§0 Kl-W oc UJ

OCR for page 46
66 a 1 oR s LU oe

OCR for page 46
67 o: => (S

OCR for page 46
68 CVJ ce. CD

OCR for page 46
69 O CO a .5 x | CO a I III CD

OCR for page 46
70 -s W) c + £ o- 3 H Q IH o 9 cr Ul Ik i z ui O ui 1818 III IU 5 2 0 I i &** l«2* O HUJ Z t_ tjt *** * oc UI Q UJ i CD Q. U- 01 a; O u. > O X S > J* c do Q « O LU oe

OCR for page 46
71 c I ta § LD ^H LU Ul

OCR for page 46
72 M >» i 111 11 > KPN •- »• 333 j Ow V «• HI SE | i - - Mijj?i il ojfa 82t:""322§8 Js x¥ = S S