A PHYLOGENETIC STUDY OF THE SHRIMP GENUS Palaemonetes HeUer 1869 FROM
NORTH AMERICA (CRUSTACEA: DECAPODA)
by
JAMES THOMAS COLLINS, B.B.A., M.S.
A DISSERTATION
IN
ZOOLOGY
Submitted to the Graduate Faculty of Texas Tech University in Partial Fulfillment of the Requirements for the Degree of
DOCTOR OF PHILOSOPHY
Approved
August, 1998 T "i ACKNOWLEDGMENTS
V J j ^ I would like to express my sincere gratitude to Dr. Marilyn A. Houck. As r ^ n chair of my committee, she provided guidance, encouragement, and enthusiasm ' throughout the course of my work in her lab. And also to Dr. Houck, a very special thank you for the long hours spent helping me in the preparation of this manuscript. I would like to thank Dr. Ned E. Strenth for introducing me to the genus Palaemonetes, and for his help and guidance over the years as I struggled to understand this group of decapods. To Dr. Richard E. Strauss, thank you for the computer programs necessary for the analysis of my data. I also wish to thank Dr. Llewellyn D. Densmore EQ and Dr. Michael R. Willig, both of whom provided support and guidance through the course of this study. I would like to express my appreciation to N. E. Strenth, H. L. McCutchen, and S. Jasper for the loan of specimens fi^om then- personal collections. I would also like to thank R. Manning for arranging for the visit to the Smithsonian and sending the specimens to Texas Tech University. I thank G. Longley for permission to collect fi"om the artesian well on campus at Southwest Texas State University. The Biology Department provided me with financial support, as a teaching assistant and a summer mini grant, for which I am extremely grateful. I would also like to acknowledge Drs. Clifford B. Fedler and Nick C. Parker and Marilyn A. Houck for allowing me the opportunity to work as a research assistant under a Department of Interior grant. It is a pleasure to acknowledge the many friends and graduate students, both in the lab and in the Department of Biology, who offered their support. A special thank you goes to Sara, Doug, Leslie, Darin, and Elizabeth who over the years have listened and offered support when I needed it. A special thank you to my wife, Gloria, who lovingly endured the years of graduate school and was instrumental in my success. I would also like to express my thanks to our son. Cliff, for encouragement and friendship along the way. And to the
ii memory of my Mother, for her support and encouragement to go forward and achieve my goal I dedicate this dissertation.
ui nn^
TABLE OF CONTENTS
ACKNOWLEDGMENTS ii LIST OF TABLES v LIST OF FIGURES vi CHAPTER L INTRODUCTION 1 Taxonomic and Systematic Review 1 Research Objectives 10 IL MATERL\LS AND METHODS 12 Specimens 12 Equipment and Handling Procedures 16 Statistical Protocols 17 Statistical Procedures 19 General procedures 19 Influence of gender on character discrimination 21 Comparison of species within the genus Palaemonetes 22 Comparison within species of P. kadiakensis 22 Evaluation of character contribution to discrimination 22 m. RESULTS AND DISCUSSION 223 Influence of Gender on Character Discrimination 223 Analysis of All Taxa 223 Comparison of Species within the Genus Palaemonetes 47 Comparison of Discrete Characters 49 Comparison within Species of P. kadiakensis 68 IV. CONCLUSIONS 91 LITERATURE CITED 94 APPENDIX: COLLECTION NUMBERS AND RAW DATA 100
IV y
LIST OF TABLES
2.1 Populations collection data 13 2.2 List of characters used in the analyses 18 3.1 Rao's V hierarchy for the 26 characters used in the analyses of all taxa 40
3.2 Rao's V hierarchy for the 30 characters used in the analyses of Palaemonetes 62
3.3 Rao's V hierarchy for the 30 characters used in the analyses of Palaemonetes kadiakensis 87
A. 1 Collection numbers and raw data utilized in this study ofthe North American Palaemonetes 100 LIST OF HGURES
3.1 Convex hulls and centroids indicating the discrimination among males and females, for all taxa studied 24
3.2 Relationship between the firsttw o axes of discrimination as determined by principal component analysis (PCA) 25
3.3 Centroids and one standard deviation about the centroids as determined by principal component analysis (PCA) for all taxa studied 26
3.4 Convex hulls and centroids indicating discrimination due to the two major shape axes as determined by principal component analysis (PCA) for all taxa studied 28
3.5 Centroids and one standard deviation about the centroids due to the two major shape axes, as determined by principal conponent analysis (PCA), for all taxa studied 29
3.6 Vector plot showing character correlations contributing to character variation resulting in shape discrimination, as determined by principal con^nent analysis (PCA), for species of all taxa studied 30
3.7 Vector plot showing character correlations representing major character variation contributing to discrimination, as determined by principal conqwnent analysis (PCA), for all taxa studied 31
3.8 Plot ofthe major size axis versus the major shq)e axis, as determined by discriminant fimction analysis (DFA) for all taxa studied 32
3.9 Centroids and one standard deviation about the centroids as determined by discriminant function analysis (DFA) for all taxa studied 34
VI 3.10 Plot ofthe major size axes, as determined by discriminant function analysis (DFA) for all taxa studied 35
3.11 Vector plot showing character correlations contributing to character variation resuhing in discrimination, as determined by discriminant function analysis (DFA), for all taxa studied 36
3.12 Vector plot showing character correlations contributing to character variation resulting in shape discrimination, as determined by discrimmant function (DFA), for all taxa studied 37
3.13 Plot ofthe cumulative Rao's V values for all taxa studied 38
3.14 Cumulative Rao's V values for all taxa studied 39
3.15 Plot ofthe major axes, as determined by size-fi^ee discriminant function analysis (DFA) for all taxa studied 41
3.16 Centroids and one standard deviation about the centroids as determined by size-fi-ee discriminant analysis (SF) for all taxa studied 42
3.17 Unrooted phenogram (UPGMA) ofthe relationships among all taxa studied, using size-fi*ee Mahalanobis distances 44
3.18 Unrooted neighbor joining phenogram of the relationships among all taxa studied 45
3.19 Neighbor joining phenogram of the relationships among all taxa studied, rooted with Leander as the outgroup 46
3.20 Convex hulls and centroids indicating the discrimination among males and females, for species of Palaemonetes 48
3.21 Relationship between the first two axes of discrimmation as determined by principal component analysis (PCA) 49
vn 3.22 Centroids and one standard deviation about the centroids as determined by principal component analysis (PCA) for species of Palaemonetes studied 50
3.23 Convex hulls and centroids indicating the discrimination (PCA) among species of Palaemonetes, due to the two major shape axes 51
3.24 Vector plot showing character correlations contributing to character variation resulting in discrimination, as determined by principal component analysis (PCA), for species of Palaemonetes studied 53
3.25 Plot ofthe major size axis versus the major shape axis, as determined by discriminant function analysis for all species of Palaemonetes 54
3.26 Vector plot showing character correlations contributing to character variation resulting in discrimination, as determined by discriminant fimction analysis (DFA), for species of Palaemonetes studied 55
3.27 Plot ofthe major size axes, as determined by discriminant function analysis (DFA) for populations of Palaemonetes studied 56
3.28 Centroids and one standard deviation about the centroids as determined by discriminant analysis for species of Palaemonetes studied 57
3.29 Vector plot showing character correlations contributing to character variation resulting in shape discrimination, as determined by discriminant function analysis (DFA), for species of Palaemonetes studied 58
3.30 Plot ofthe cumulative Rao's V values for all species of Palaemonetes studied 60
3.31 Cumulative Rao's V values for all species of Palaemonetes studied 61
Vlll 3.32 Plot ofthe major axes, as determined by size-fi-ee discriminant function analysis (SF) for all species of Palaemonetes studied.. 63
3.33 Centroids and one standard deviation about the centroids for the shape axes, as determined by size-fi-ee discriminant analysis (SF), for species of Palaemonetes studied 64
3.34 Unrooted neighbor joining phenogram ofthe relationships among species of Palaemonetes studied 65
3.35 Unrooted phenogram (UPGMA) ofthe relationships among Palaemonetes species studied, using size fi-ee Mahalanobis distances 66
3.36 Convex hulls and centroids indicating the discrimination among males and females, for populations of P. kadiakensis 69
3.37 Relationship between the firsttw o axes of discrimination as determined by principal component analysis (PCA); convex hulls with centroids for populations of P. kadiakensis studied 70
3.38 Centroids and one standard deviation about the centroids as determined by principal component analysis (PCA) for populations of P. kadiakensis studied 71
3.39 Vector plot showing character correlations contributing to character variation resulting in shape discrimination, as determined by discriminant function analysis (DFA), for species of Palaemonetes studied 73
3.40 Convex hulls and centroids indicating the discrimination among all taxa (PCA), due to the two major shape axes for P. kadiakensis populations studied 74
3.41 Centroids and one standard deviation about the centroids due to the two major shape axes, as determined by principal component analysis (PCA), for populations of P. kadiakensis studied 75
IX 3.42 Vector plot showing character correlations contributing to character variation resuhing in shape discrimination, as determined by principal component analysis (PCA), for the populations of P. kadiakensis studied 76
3.43 Plot ofthe major size axis versus the major shape axis, as determined by discriminant function analysis (DFA) for populations of P. kadiakensis studied 77
3.44 Plot ofthe major size axes, as determined by discriminant function analysis (DFA) for populations of P. kadiakensis studied 78
3.45 Vector plot showing character correlations contributing to character variation resuhing in shape discrimination, as determined by discriminant fimction analysis (DFA), for the populations of P. kadiakensis studied 79
3.46 Vector plot showing character correlations contributing to character variation resulting in shape discrimination, as determined by discriminant function analysis (DFA), for the populations of P. kadiakensis studied 80
3.47 Plot ofthe major axes, as determined by size-fi-ee discriminant function analysis (SF) for all populations of P. kadiakensis studied 81
3.48 Vector plot showing character correlations contributing to character variation resulting in shape discrimination, as determined by size-fi-ee discriminant analysis (SF), for the populations of P. kadiakensis studied 82
3.49 Centroids and one standard deviation about the centroids for the shape axes, as determined by size-fi-ee discriminant analysis (SF), for populations of P. kadiakensis studied 84
3.50 Cumulative Rao's V values for populations of P. kadiakensis studied 85
3.51 Plot ofthe cumulative Rao's V values for P. kadiakensis 86 3.52 Unrooted phenogram (UPGMA) ofthe relationships among Palaemonetes kadiakensis populations using size-fi-ee Mahalanobis distances 88
3.53 Unrooted neighbor joining phenogram ofthe relationships among populations of Palaemonetes kadiakensis, given by county of collection 90
XI CHAPTER I INTRODUCTION
Taxonomic and Systematic Review The genus Palaemonetes Heller, 1869, was first described as Palaemon varians from the eastern Atlantic by (Leach 1814) (Holthuis, 1952). Since that time, several species of Palaemonetes have been described from all ofthe continents with the exception of Antarctica. In North America there are currently 14 described species (Palaemonetes vulgaris (Say, 1818); P. paludosus (Gibbs, 1848); P. antrorum Benedict 1896; P. kadiakensis Rathbun, 1902; P. hiltoni Schmitt 1916; P. intermedins Holthuis, 1949; P. pugio Holthuis, 1949; P. cummingi Chace, 1954; P. suttkusi Smalley, 1964; P. lindsqyi Villalobos and Hobbs, 1974; P. texanus Strenth, 1976; P. holthuisi Strenth, 1976; P. mexicanus Strenth, 1976; and P. hobbsi Strenth, 1994). Of these, 13 species occur in the eastern United States and Mexico, and only P. hiltoni is known fi-om the southwest coast ofthe United States and the northwest coast of Mexico (Fig. 1.1). Although these species of Palaemonetes are an important part ofthe temperate and tropical aquatic food webs (Lowe and Provenzano, 1990; Garcia, 1991), they have little conmiercial value (Strenth, 1976). In North America, their value has been restricted to use as commercial fish food (Worth, 1908; Nielson and Reynolds, 1977); however, in Viet Nam, P. camranhi, Palaemon semmelinkii and Periclimenes grandis are used as a dietary supplement for humans (Xuan, 1997). The lack of commercial importance, the occurrence of morphological homogeneity among taxa (Strenth, 1976), and the lack of "understanding ofthe currently used taxonomic characters" (Strenth, 1976, p. 1) contribute to the lack of resolution ofthe systematic relationships for this group of natant decapods. Two ofthe freshwater epigean species, P. paludosus and P. kadiakensis, range over a wide area. Palaemonetes paludosus is found primarily along the east coast ofthe United States fi-omFlorid a to New Jersey east ofthe Allegheny ^
CO CO :a
CO 0) • 1-H CO fi
'o o
fi o :3 u CO
be Mountains (Holthuis, 1952). However, several authors have reported populations of P. paludosus ranging across the southern United States as far west as Texas (Creaser and Ortenburger, 1933; Meehean, 1936; Holthuis, 1952; Webb, 1980; Garcia, 1991). Palaemonetes kadiakensis ranges fi-om northeast Mexico, throughout the Mississippi River Valley, and into Wisconsin and Michigan (Creaser, 1932; Nelson, 1982; Hobbs and Jass, 1988). The other fi-eshwater epigean species exhibit restricted habitats. Palaemonetes lindsayi, P. suttkusi, P. mexicanus, P. hobbsi, and P. texanus, are artesian spring forms only found within a few hundred meters of spring discharges. Subterranean species also exhibit a fairly restricted distribution. Two ofthe subterranean species, P. antrorum and P. holthuisi, are known only to occur in the Edwards Aquifer of central Texas. The type locality for P. antrorum is an artesian well on the campus of Southwest Texas State University in San Marcos, Texas. However, it has been reported fi-omth e Edwards Aquifer (Ulenhuth, 1921) as far west as Uvalde County, Texas (Hobbs et al, 1977). Palaemonetes holthuisi also is found in the Edwards Aquifer. It was described fi-omEzell' s Cave (Strenth, 1976) which is only a few miles west ofthe type locality for P. antrorum. Ulenhuth (1921) beUeved that both the artesian well at Southwest Texas State University and Ezell's cave are part ofthe subterranean Purgatory Creek System. Webb (1980) reported a single individual of P. holthuisi collected fi-omth e artesian well on the campus of Southwest Texas State University. Garcia (1991) also reported having found a smgle individual of P. holthuisi but fails to give the location. The marme or estuarine species, P. vulgaris, P. intermedins, and P. pugio, have wide geographic distributions. Holthuis (1952) in his revision ofthe family Palaemonidae, noted that all three species range along the east and south coasts of the United States fi"omMassachusett s to Texas. Palaemonetes vulgaris generally inhabits brackish and salt water, while P. pugio occurs in brackish to ahnost fi-eshwater (Hohhuis, 1952). The smgle marine or estuarine species fi-omth e west coast of North America, P. hiltoni, ranges from southern California to the Mexican states of Sonora and Sinaloa. According to Holthuis (1952), little is known about the habitat of this species; however, he notes that the specimens collected from Del Mar were collected in a slough and those from Sinaloa were from an estuary. When Holthuis (1949; 1952) revised the genus Palaemonetes, his classification and taxonomic descriptions focused on the identification ofthe currently described species rather than on their evolutionary relationships. Since that time various authors have produced studies on population genetics and systematics on some species of Palaemonetes fi-omNort h America (Webb, 1980; Collins, 1993; Garcia and Davis, 1994). Pereira (1989) examined the superfamily Palaemonoidea and proposed systematic relationships above the generic level. Although these studies have provided information about the taxonomy and systematics of this group, the systematic relationships among species ofthe genus Palaemonetes fi-omNort h America remains unclear. Based on egg size, Sollaud (1923) divided the family Palaemoninae into two groups, one with large numbers of small eggs and the other with fewer numbers of large eggs. During an examination ofthe different species listed in the revision by Holthuis (1950), Strenth (1976) found that marine species exhibited numerous small eggs and freshwater species exhibited fewer but larger eggs, and concluded that egg number and size substantiate the separation of Palaemonetes into two groups, as first proposed by Sollaud (1923). Numerous studies (Broad, 1957; Broad and Hubschman, 1962, 1963; Dobkin, 1963, 1971; Hubschman, 1974; Strenth and Longley, 1990, Strenth, 1991) have documented developmental differences among taxa based on the number of larval stages. Freshwater species with larger eggs generally have fewer larval stages whereas marine forms exhibit more larval stages. Palaemonetes kadiakensis is the exception to this rule in that its development more closely resembles that of marine taxa (Broad and Hubschman, 1963). Generally, the freshwater and the marine groups can be separated on qualitative morphological criteria. Holthuis (1949, 1952) noted that there was a difference in the number of fused segments and the length ofthe unfused distal end ofthe upper antennular flagellum. Freshwater species, with the exception of P. antrorum and P. holthuisi (Strenth, 1976) have the free part of their antennule shorter than the fused part. In marine species and the fi-eshwaterspecie s P. antrorum and P. holthuisi the free distal end is longer than the fused portion (Holthuis, 1949, 1952; Villalobos et al., 1974; Strenth, 1976). This similarity between antennule morphology might be attributed to an adaptation to low light levels by the subterranean species that are tactility oriented and have degenerate eyes. Generally, a difference exists in the anterma morphology ofthe form I zoea stage of freshwater and marine Palaemonetes. With one exception, P. kadiakensis, the North American freshwater species of Palaemonetes exhibit an unsegmented antenna scale, whereas marine species have a segmented antenna scale in their form I zoea. Differences in egg number and size, antennule morphology, and development between fi-eshwater and marine groups, strongly supports the monophyly ofthe fi-eshwater species of Palaemonetes from North America. The evidence examined by Strenth (1976) led him to conclude that the freshwater species of Palaemonetes fi-om North America are in fact a monophyletic group. Since Holthuis' work (1949, 1952), there have been several attempts to explain the systematic relationships within Palaemonetes. In his revision ofthe species of Palaemonetes from North America, Holthuis (1952) placed P. antrorum in the subgenus Alaocaris. AW remaining North American species were placed in the subgenus Palaemonetes. This separation was based on four characteristics: (1) degenerated eyes without pigment, (2) absence of teeth on the lower margin ofthe rostrum, (3) similarity ofthe shape and size ofthe first and second pereiopods, and (4) lack of a movable exopod spine on the uropods. With the description of several new species of Palaemonetes from North America, the separation of P. antroum into its own subgenus became untenable and Strenth (1976) synonymized the subgenera Palaemonetes and Alaocaris. Vandel (1965) noted that it is common for eyes in cavenerincole organisms to be reduced. This is especially true of subterranean species of Palaemonetes fi-om North America (Benedict, 1896; Chace, 1954, Strenth, 1976). Strenth (1976) noted that both P. antrorum and P. holthuisi did not have teeth on the lower margin of their rostrum. The reduction of lower margin dentition is common in the subterranean species of palaemonids (Strenth, 1976). There appears to be a correlation between the reduction of lower rostral dentition and degeneration of eyes, which Strenth (1976) attributes to a "specialization associated with life in the subterranean environment" (p. 14). The shape and size ofthe legs of P. antrorum are also similar to those of P. holthuisi. These two species feed while suspended on their last three legs, using the first two pairs of pereiopods extended, rather than the crouched position typical of epigean species. Leg morphology is an adaptation for living in a subterranean environment (Strenth, 1976). At the time Holthuis (1952) revised the genus Palaemonetes there were only three fi-eshwaterspecie s described from North America. Palaemonetes antrorum differed fi-omP . paludosus and P. kadiakensis in that it lacked the movable exopod spine and was fi-om a subterranean environment. Since this revision three additional fi-eshwater species of Palaemonetes have been described that lack the moveable exopod spine (Smalley, 1964; Villalobos and Hobbs, 1974; Strenth, 1976). With the addition ofthe epigean species, which do not exhibit a movable exopod spine the validity of this character for the separation of P. antrorum into a separate subgenus is greatly reduced Although the variability ofthe movable exopod is quite high, Strenth (1976) stated that it is an important character for understanding the systematic relationships among the fi-eshwaterspecie s of this group. Based on the condition ofthe movable exopod spine and geographic distribution, he divided the fi-eshwaterspecie s into three groups. The first, consisting of P. kadiakensis, P. paludosus, P. cummingi, exhibits a movable spine on both exopods and is distributed fi-om north eastern Mexico to the east coast ofthe United States. Palaemonetes antrorum, P. holthuisi, P. suttkusi, and P. lindsayi compose the second group, in which the movable spine is absent on both exopods. This group is distributed fi-omcentra l Texas to Mexico, west ofthe Sierra Madre Oriental Mountains. The range ofthe third group extends fi-om central Texas into Mexico along the eastern side ofthe Sierra Madre Oriental Mountains. In this group, the movable exopod spine "is quite variable; both movable spines may be present, both may be absent, or only a left or right one may be present within a single population" (Strenth, 1976, p. 5). At the time Strenth divided the fi-eshwater species into these three groups, this third group consisted of P. texanus and P. mexicanus. Subsequently, P. hobbsi, was named (Strenth, 1994). This species is morphologically consistent as to the movable exopod spine. In addition, it is a geographically intermediate between P. texanus and P. mexicanus. The relationship among the three groups was examined most recently by Collins (1993) using allozyme electrophoresis. This study was restricted in scope to the species fi-omTexa s and Mexico. However, there was strong evidence to support P. texarms, P. hobbsi, and P. mexicanus as a single group. The study also indicated that P. lindsayi, P. suttkusi, and P. antrorum may indeed form a natural group. Because only P. kadiakensis was examined fi-om the group containing spines on both exopods, any possible relationship among the member species of this group could not be addressed. It should be noted; however, that P. kadiakensis was not included in either group. Further studies are needed to determine if P. kadiakensis, P. paludosus, and P. cummingi form a monophyletic group (Collins, 1993). The use of male external genitalia as a taxonomic character was proposed by Fleming (1969) in the hope that they might resolve systematic relationships within epigean species of Palaemonetes from North America. Using apical and subapical setae from the tip ofthe appendix masculina, he found that P. kadiakensis, P. paludosus, and P. pugio can be discriminated fi-omeac h other and that they can be distinguished from P. vulgaris and P. intermedins. Since Fleming's work, additional species have been described and male genitalia subsequently have been used in the description (Villalobos and Hobbs, 1974; Strenth, 1976, 1994). A study by McCutchen (1983) on the morphological variation ofthe North American fi-eshwater species of Palaemonetes found features ofthe appendix masculina did separate marine and freshwater species. However his results clearly indicate that vsathin freshwater species the variation of spines within species is greater than among species. McCutchen (1983) concluded that the appendix masculina was not a valid taxonomic character for the fi-eshwaterspecie s of Palaemonetes from North America. Until Holthuis (1949, 1950, 1952) revised the family Palaemonetes, P. paludosus was considered to be conspecific with the Palaemonetes fi-omth e west slope ofthe Alleghenies (Dobkin, 1963). However, as was pointed out above P. kadiakensis exhibits valid characteristics which discriminate it fi-omP . paludosus (form I zoea morphology and number of larval stages). According to Dobkin (1971), even though these differences in larval morphology and development occur, the adult forms appear to be more closely related to each other than to other species within this genus. Dobkin (1963, 1971) proposed that P. kadiakensis may be a recent immigrant into the fi-eshwater environment. He has noted that vdth the exception of P. argenimts (Nobili 1901) and P. kadiakensis, the known fi-eshwater species of Palaemonetes exhibit an abbreviated larval development. This he attributes to the possibility that species that have been in the freshwater longer would tend toward the abbreviated larval development. Based on this, he hypothesized that P. kadiakensis
8 may be a relatively recent immigrant into the fi-eshwaterenvironmen t (Dobkin. 1963. 1971). Palaemonetes argentinus exhibits characteristics more closely related to the marine species which lead Strenth (1976) to conclude that P. argentinus probably is related more closely to the marine species, and represents a different evolutionary pathway into fi-eshwatera t least in comparison with the fi-eshwater species fi-om North America. Based on the larval development, form I zoea antennular morphology, and similarity among the early larval stages of P. kadiakensis, P. pugio, P. vulgaris, and P. intermedins the hypothesis of Dobkin (1965, 1971) and Strenth (1991) that P. kadiakensis represents a more recent invasion into the fi-eshwater enviroimient is supported. An examination of allozymes and DNA ofthe Palaemonetes of Texas revealed two different haplotypes of P. kadiakensis based on genetic disequilibrium of selected markers (Garcia, 1991; Garcia and Davis, 1994). Populations of both haplotypes were sympatric in many locations and maintained reproductive isolation. However, hybridization occurred in several locations in central Texas (Garcia, 1991; Garcia and Davis, 1994). Garcia (1991) erected a clade ofthe Texas species of Palaemonetes based on 17 loci, (P. paludosus, P. texanus, and P. holthuisi). Palaemonetes kadiakensis was the sister group to this (Garcia, 1991). Webb (1980) re-erected the suhgeneius Alaocaris to include P. antrorum and P. holthuisi based on the fact that both species live in a subterranean habitat. His electrophoretic data suggested that the Palaemonetes fi-om Texas should be divided into three groups: (1) a marme group consisting of P. pugio, P. intermedins, and P. vulgaris; (2) a fi-eshwatergrou p (P. texanus, P. kadiakensis, P. paludosus): (3) the troglobites, P. antrorum and P. holthuisi. Separation ofthe marine and fi-eshwater species does not appear to be in question (Strenth, 1976); the separation ofthe fi-eshwater species fi-omth e trogolobites is not accepted by current researchers of this group. In a re-evaluation of P. holthuisi, Bruce (1993) elevated this species to the new genus Calathaemon. He based this on the fact that the mouthparts are radically different fi-om those of other Palaemonetes. In Palaemonetes the mouthparts are adapted primarily for scavenging and possibh' predation. The mouthparts in Calathaemon are adapted as a fihering system (Bruce, 1993). Bruce (1991) noted that the branchiostegal region forms a cavity allowing movement ofthe caridean lobe and the scaphognathite. These movements may produce strong currents capable of drawing food particles across the filtering mechanisms. Bruce (1991) notes that this process has not actually been observed. Strenth (1976, pers. comm.) indicated the feeding habits of both P. antrorum and P. holthuisi are quite similar and not unlike that ofthe epigean species of Palaemonetes. Bruce (1991) reported that the chelae on both the first and second pereiopods was distinctively different fi-om that of other members ofthe family Palaemonidae. The central region of both chelae appear flattened or possible convex in Calathaemon, rather than concave as in other species of Palaemonetes. He also noted that the chelae on both pairs of pereiopods of Calathaemon are similar rather than distinct as in species of Palaemonetes.
Research Objectives The broad research objective of this study is to understand the systematic relationships ofthe species within the genus Palaemonetes fi-omNort h America. This will provide a fi-amework to increase our knowledge of this important part of the aquatic food chain, as well as provide a mechanism for understanduig more about the biogeography and the ecology of North America. The specific issues to be examined in this study are: (1) the development of hypotheses of systematic relationships among species of Palaemonetes from North America; (2) whether the fi-eshwater species do form a monophyletic group as proposed by Sollaud (1923) and Strenth (1976); (3) whether following the hypothesis by Holthuis (1952), and re-enacted by Webb (1980), P. antrorum should be elevated to a subgenus {Alaocaris) and separated fi-omth e other species of North American Palaemonetes; (4) whether P. holthuisi should be elevated to a new genus.
10 Calathaemon, as proposed by Bruce (1993) is supported by the data; (5) whether the hypothesis by Dobkin (1965, 1971) that P. kadiakensis is a recent immigrant into the fi-eshwater environment is upheld; (6) whether the origin ofthe North American species of Palaemonetes was in sah of fi-eshwater; and (7) what the potential role of biogeography in the evolution ofthe genus Palaemonetes.
11 CHAPTER II MATERIALS AND METHODS
Specimens The focus of this research was the discrimination among species of Palaemonetes fi-om North America. However, to put this taxon in perspective relative to other genera within the family Palaemonidae, additional taxa were referenced to determine the appropriate outgroup comparison. The genera considered to be potential out-group taxa were: Leander, Macrobrachium. Palaemon and Calathaemon. Specimens which represented the genus Leander (L. tenuicornis) were collected m Galveston, Texas, and were borrowed for this study fi-omN . E. Strenth of Angelo State University. Macrobrachium {M. acanthums) representatives were collected fi-om Montego Bay, Jamaica and Palaemon (P. northropi) was collected fi-om Guadalupe Island in the Caribbean. Calathaemon (=Palaemonetes holthuisi), Macrobrachium and Palaemon were borrowed fi-om the National Museum of Natural History (NMNH). Specimens of Palaemonetes used in this research were: collected by the author, loaned by individuals that shared their personal collections, or borrowed fi-om the NMNH. The type specimens of two species, Palaemonetes cummingi and P. holthuisi were examined on location at the NMNH. There were a total of 19 populations representing 14 species of Palaemonetes fi-om North America examined in this study. All nominal species of North American Palaemonetes were included in this research project. The collection locations for each species and population, number of mdividuals used in this study, and number of males/females are Usted in Table 2.1.
12 Table 2.1. Population collection data including taxa, total number of individuals used in this study and the number of males (m) and females (f) and the collection location. Species Total No. (m/f) Location Palaemonetes P. antrorum 10(5/5) Artesian well on the campus of Southwest State University campus, San Marcos, Hays Co., Texas. This is the type locality for this species. P. cummingi 4(1/3) Squirrel Chimney, Alachua Co., Florida. This is the type locality for this species. The specimens were examined in the NMNH. P. hiltoni 9 (4/5) On loan fi-omth e NMNH. Collected in the Mexican state of Sinaloa. P. hobbsi 10(5/5) The Nacimiento de Rio Mante, Tamaulipas, Mexico. This is the type locality for this speces. P. holthuisi 2 (2/0) Ezell's Cave, Hays Co., Texas, {Calathaemon) type locality for this species. P. intermedins 8 (5/3) On loan fi-omth e NMNH. Collected fi-omLm k Port, St. Lucie Co., Florida. P. kadiakensis 9 (4/5) Sapo Lake, Hidalgo, Co., Texas 4 (2/2) Eagle Nest Canyon, Langtry, Val Verde Co., Texas 6 (1/5) Highway 36, N. of Caldwell, Burleson Co., Texas. 5 (1/4) Mullins Crossmg, Tom Green Co., Texas.
13 Table 2.1. Continued
Species Total No. (m/f) Location P. kadiakensis 10(6/4) San Saba River, Menard, Menard Co., Texas. 3(2/1) Colorado River at the Highway 29 bridge, Llano Co., Texas. P. lindsayi 10(5/5) 15 miles W. of Ciudad Valles, San Luis Potosi, Mexico. This is the type locality for this species. P. paludosus 10(5/5) Ashley River, W. of Charleston, South Carolina. Thought to be about 10 miles up stream fi-om the type locality. P. pugio 9 (4/5) Goose Island State Park, Aransas Bay, Texas. P. suttkusi 10 (5/5) Cuatro Cienigas Basin, Mexico. This is the type locality for this species. P. texanus 10(5/5) San Marcos River, San Marcos, Hayes Co., Texas. This is the type locality for this species. P. vulgaris 9 (5/4) Skidaway River estuary. University of Georgia Aquarium, Skidaway Island, Savaimah, Georgia. Leander tenuicornis 10(0/10) Galveston, Texas. Macrobrachium acanthums 4(1/3) On loan fi-omth e NMNH. Collected fi-om a brackish pond, Montego Bay, Jamaica.
14 Table 2.1. Continued
Species Total No. (m/f) Location Palaemon northropi 10(5/5) On loan fi-om the NMNH. Collections site was Caribbean Sea, Point A Pitre Guadalupe Island.
15 Epigean specimens collected by the author were captured in relatively shallow water with a hand dip net. Specimens were collected as close as possible to the type locality for each species. All were collected either from aquatic vegetation or from under overhangs ofthe bank. The subterranean species P. antrorum was collected by securing a net over the artesian well discharge on the Southwest Texas State University in San Marcos, Texas. Once specimens were collected, they were placed in separate vials containing 70% ETOH. Each individual was given an identification number and the location of the collection site was recorded. The samples of populations provided by other collectors or the NMNH also were preserved in 70% ETOH at the time of their collection and permanently stored in alcohol. Individual specimens were chosen for this study based upon the quality of their physical condition. Specimens that were lacking appendages were rejected, when possible, but retained in damaged type specimens. Also, an attempt was made to select an equal number of males and females from randomly sampled populations. The NMNH collection numbers for the type species were retained and used by the author (Appendix). All others specimens fi-om NMNH arrived at Texas Tech University as bulk-collections. Subsets of these bulk-collections were parceled into individual vials for study. The vials were then individually numbered and those numbers were retained (in addition to the bulk identification numbers) for future reference when the specimens were returned to the NMNH.
Equipment and Handling Procedures Specimens were removed from the ETOH and placed in a dish containing wax so individuals could be piimed in the appropriate orientation to facilitate measurement. Specimens were measured under an Olympus SZHIO research stereo microscope with zoom capabilities l-70x. Images were displayed on a video monitor using a Sony CCD/RGB color camera connected to a Sony Trinitron (PVM1341) monitor. The images were then captured using Image-Pro Plus (V.2)
16 software fi-om Media Cybernetics, Maryland. All ofthe specimens were measured directly fi-omth e image on the monitor. Measurements were consistently taken fi-om the right side ofthe specimen unless that particular character was damaged (e.g., type specimens). Because the size range of individuals differed, it was necessary to image various specimens at different relative magnifications. Character measurements (Table 2.2) were then converted to millimeters, based on a calibrated constant. The constant was calibrated by repeated measurements of a known distance at each magnification. The constant was then applied to the raw measurement. It was necessary to measure both P. cummingi and P. holthuisi on site at the NMNH. In this case, the specimens were measured using Sylvac Ultra-Call Mark III calipers manufactured by Fred V. Fowler, Co. Only linear measurements could be taken in this maimer on these specimens. To test the comparability ofthe caliper measurements with others, caUper measurements also were made on mdividuals fi-om the several other species, and compared to the same measurement taken fi-om the captured image. The distance measurements were comparable (varied by no more than 2%).
Statistical Protocols Anatomical landmarks and helping points {sensu Bookstein et al, 1985) were selected to defme characters used for this study. Landmarks provided information about homologous points that show anatomical delineation. Helpmg points are non homologous points that aid in defming morphological surfaces, when recorded m association with landmarks. There mitially were a total of 40 distance characters mcluded in this study. However, the number of characters was reduced to 30 (Table 2.2) because of missing data among a subset of individuals. The EucHdean distances for the 30 characters were calculated among the landmarks for 172 individuals in order to assess character
17 Table 2.2. List of characters used in the analyses. Characters are referenced by number m the vector plots. Allometric coefficients relate to the principal component analyses of all taxa. Character # Description of Distance Characters AJlometries 1. Rostrum length, from tip of rostrum to base of last 0.9664 tooth along the dorsal edge 2. Carapace Length, from tip of rostrum, along the 1.1565 dorsal margin, to the posterior edge ofthe cephalothorax 3. Ischium Leg 1, from apodeme to apodeme 1.0805 4. Merus Leg I, from apodeme to apodeme 1.1591 5. Corpus Leg I, from apodeme to apodeme 1.0646 6. P*ropodus Leg I, from apodeme to apodeme 0.9861 7. Dactyl Leg 1, from apodeme to apodeme 0.8940 8. Ischium Leg II, from apodeme to apodeme 0.9392 9. Corpus Leg II, from apodeme to apodeme 1.0974 10. Corpus Leg II. from apodeme to apodeme 0.9825 11. Propodus Leg II, from apodeme to apodeme 0.9934 12. Dactyl Leg II, from apodeme to apodeme 0.9462 13. Ischium Leg IV, from apodeme to apodeme 1.0389 14. Merus Leg IV, from apodeme to apodeme 1.1195 15. Corpus Leg IV, from apodeme to apodeme 0.9540 16. Propodus Leg IV, from apodeme to apodeme 1.1107 17. Dactyl Leg IV, from apodeme to apodeme 1.0072 18. Ischium Leg V, from apodeme to apodeme 1.0082 19. Merus Leg V, from apodeme to apodeme 1.0920 20. Corpus Leg V, from apodeme to apodeme 0.8915 21. Propodus Leg V, from apodeme to apodeme 1.0270 22. 3rd Maxili Carpus, from apodeme to apodeme 1.0830 23. 3rd Maxili Propodus/Dactyl, from apodeme to 0.8732 apodeme 24. 2nd Segment Antenna Peduncle, from apodeme to 0.8606 apodeme 25. 3rd Segment Antenna Peduncle, from apodeme to 0.7762 apodeme 26. Telson M x M. a line aaoss the telson from margin 1.1089 to margin, located at the base ofthe anterior spines 27. Telson L x L, a line across the telson from margin to 1.0325 margin, located at the base ofthe posterior spines 28. Telson M x L, between the bilateral lines formed at 0.9972 the base ofthe anterior and posterior spines 29. Telson B X M, from the anterior base ofthe telson to 0.9465 the bilateral line aaoss the telson at the anterior spines 30. Telson L x T. from the bilateral line at the base ofthe 0.8069 posterior pair of spines to the tip ofthe telson
18 variation. An attempt was made to have an equal number of males and females for each species or population (see Appendix). The distances measured were chosen so as to be homologous among inidviduals and populations. Characters also were chosen on the criteria of repeatability. Only the characters that were reasonably fi-ee of distortion (due to handling) were measured. Characters were distributed across all elements ofthe animal form. Quantitative phyletic taxonomy as proposed by Kluge and Farris (1969) does utilize biological information and can be used to discover systematic and evolutionary relationships among organism. Data of a continuous nature, immunological, molecular are quantitative and must be analyzed using numerical techniques (Farris, 1972). Felsenstein (1984) noted phylogenetic distance methods are independent ofthe clustering technique used. Colless (1970) concluded that phylogenies of contemporaneous species could be reconstructed using phenetic methods. Because morphological distance measurements are continuous data, analyses using numerical techniques is appropriate.
Statistical Procedures General procedures All ofthe following statistical procedures were performed consistently across groups (Table 2.1) analyzed simultaneously (or as subsets), as the research questions demanded. However, specific statistical adjustments were required for the within- Palaemonetes comparison and the within-P. kadiakensis analyses (detailed below). In order to preserve allometric scaling effects among characters, alleviate effects of heteroscedasticity (Bookstein et al., 1985), and produce a scale-invariant covariance matrix, all character distances were converted to natural logarithms (Jolicoeur, 1963) prior to fiirther analysis.
19 Principal components analysis (PCA) was preformed on a covariance matrix of log-transformed data to determine possible structural relationships among the variables (Bookstein et al., 1985; Digby and Kempton, 1987). Individuals, or groups of individuals, were not assigned to any a priori categorical group, prior to running the PCA. And, the variation of all 30 characters was examined simultaneously for all taxa. Estimates ofthe relative allometric coefficients of individual characters, with reference to the general size-vector (PCI) were produced (Table 2.2). The populations were then subjected to discriminant fimction analysis (DFA). This procedure minimizes variation within designated groups and maximizes the variation between groups (Wiley, 1981). In order to discriminant groups on the basis of shape rather than size, size was regressed out ofthe data set prior to running a second DFA. Mahalanobis distances (D^) were calculated among groups, in the full multivariate space. Size was partitioned out ofthe data set by regressing each log-transformed character on PCI, which is typically the major size axis for biological forms. This provided the ability to do a size-fi-ee analysis (SF) and allowed the biological patterns in form to be interpreted as size-fi-ee shape contrasts. Centroids and confidence ellipses were generated by group to simplify visualization ofthe PCA plots, the DFA plots, and the central tendency ofthe distributions. The character correlations were plotted as vector diagrams (Wright, 1954; Strauss, 1985). Because DFA coefficients (weights) are not necessarily biologically interpretable it was necessary to do the vector analysis in this way. Phenetic clustering ofthe taxonomic groups was performed on Mahalanobis distances using a clustering algorithm (UPGMA) ofthe size-free DFA. This procedure allowed an interpretation ofthe distance relationships among all
20 characters in the full muhivariate space, assuming equal rates of evolution for the characters. To evaluate taxonomic relationships, releasing the constraint of equal rates of evolution, additive dichotomous trees (neighbor-joining dendrograms) were constructed to explore Leander, Macrobrachium, Palaemon and Calathaemon as outgroups. All statistical analyses were performed on an IBM PC using commercial Matlab® (1992) routines and Matlab® routines (m files) generously provided by R. E. Strauss. The PCA, DFA, SF, vector plots, allometries, distance phenogram diagrams, and neighbor-joining trees were all produced using Matlab® m files.
Influence of gender on character discrimination Sexual dimorphism in the full multivariate character space was assessed to determine whether males and females could be pooled, within taxa and populations, for further analysis. However, due to small sample size of some groups (e.g., type specimens), sexual dimorphism within species and within taxa could not be statistically tested using a MANOVA. However, there is indirect inference that does not allow the rejection ofthe null hypothesis that males and females represent a common population in multivariate character space. The direct inference extends from that fact that: (1) a one-way MANOVA on a testable subset of organisms (i.e., P. kadaikensis) did not indicate any sexual dimorphism, (2) active researchers investigating the Palaemonidae have not uncovered any documented dimorphism in taxa studied (pers. comm., N. Strenth). It is a reasonable, yet an untestable inference that there is no significant sexual dimorphism among study organisms used in this work, but the question warrants further separate consideration.
21 Comparison of species within the genus Palaemonetes The genus Palaemonetes was examined as a single taxon, assuming the monophyly of relationships of North American species supported in recent literature. However, P. antrorum was found in this study to be significantly divergent fi-om all others, and a separate series of analyses were run after removing P. antrorum, in some cases. This was done in order to further define the internal relationships among the remaining taxa.
Comparison within species ofP kadiakensis Discrimination and phenetic relationships within species P. kadiakensis was not possible on the fiill data set because the number of characters exceeded the number of individuals. Character sorting was necessay to determine those characters which were most influential in character discrimination and such characters were retained for the muhivariate analyses. Characters were evaluated using Rao's V, to determine which characters contributed most to the discrimination within the group (Tabachnick and Fidell, 1989). The Rao's V procedure resuhed in a final character set reduction to 30 distance characters (and for all taxa, 2) which was entered into a DFA. Multivariate static allometric coefficients (Om) were scaled with respect to PCI by estimating the loadings (coefficients) on PCI for each character and scaling the loadings to a mean of 1.0 (Houck et al, 1990; Houck 1992; Hutcheson et al. 1995; Houck and OConnor, 1998).
Evaluation of character contribution to discrimination The Rao's V procedure was applied to: all taxa, all species of Palaemonetes (excluding P. antrorum), and to P. kadiakensis only. This was done to evaluate the relative importance of characters at various taxonomic levels.
22 CHAPTER m RESULTS AND DISCUSSION
Influence of Gender on Character Discrimination An initial principal component analysis (PCA) of characters for males separate fi-om females was completed on all taxa studied in order to determine whether there was character variation that could be attributable to sexual dimorphism. The convex hulls ofthe PCA, by gender, exhibited a significant amount of overlap, with the centroids relatively close together (Fig. 3.1). The centroids on PCI differed by approximately one standard deviation (SD) unit in the full space, and on PC2 by approximately 0.25 SD units. This indicates that there is little or no evidence for significant sexual dimorphism in this sample of organisms and thus data from males and females was pooled in subsequent analyses. A MANOVA was attempted, to establish an F statistic for gender differences among all taxa in the full space, but the individual sample sizes ofthe taxa were not large enough to contribute to a meaningfiil comparison.
Analysis of All Taxa Examining all taxa studied, the first principal component (PCI) accounted for 67.5% ofthe total variation in the data set. PCI is a significant size vector as indicated by the magnitude and direction ofthe loadings and an apparent size trajectory that is correlated with the first eigenvector (Fig. 3.2). The plot of PC2 versus PCI suggested that Leander is the apparent outgroup ofthe taxa studied, with the genersi Macrobrachium, Palaemon and Calathaemon positioned well within a cluster ofthe remaining taxa (Fig. 3.2). Separation of groups in multivariate PCA space is more clearly visualized by the centriod plot (Fig. 3.3). The centroid analysis indicated that Macrobrachium is the largest ofthe taxa examined and P. antrorum is this smallest. This is
23 All Taxa ^ ^,'' \ ^r ^ , " ' \^ \ ^r » " "" ^v \ ^r ^ " ^ ^S. \ / '' \ « X," x /^ / Males \ •' / •' * • Females ^'^ v o (N / ' ^-"-^^ u / ^^—""""""''^ ' OM
1 1 1 " ^ / '
" -» / ^— ^ ^
PCI (67.5%)
Figure 3.1. Convex hulls and centroids indicating the discrimination among males and females, for all taxa studied. PCI, the major size axis, accounts for 67.5% ofthe variation in all taxa examined, and PC2, a major shape axis, accounts for 10.3%. PCI ranges from 0.0-6.0 standard deviation units, and PC2 ranges from 0.5-1.5 standard deviation units.
24 All Taxa Calathaemon /y^Z^^^^
Freshwater >^^^5^^yf • X -"' -•^^^Vc^ Macrobrachium
©^ m ^^^^^^^—^^ ^# ^ Ul^ ^^^^^fc^^^^^^ >(10 .
-^^^ • J .^^ CL, .^^^ \ * ^-.-"^"'''''^^^^O^'^''''^
Marine
y^^^^-^^^Leander C • ^*'^^^^^
PCI (67.5%) Figure 3.2. Relationship between the first two axes of discrimination as determined by principal component analysis; convex hulls and centroids for all taxa studied. The marine grouping consists of Leander, Macrobrachium, Palaemon, Palaemonetes hiltoni, P. intermedins, P. pugio, P. vulgaris plus P. paludosus and P. suttkusi which are freshwater inhabitants. The freshwater taxa consist of Calathaemon, P. antrorum, P. cummingi, P. hobbsi, P. kadiakensis (six populations), P. lindsayi, and P. texanus. The dotted line indicates a separation ofthe basically fresh and salt water groups. The bold convex hulls represent genera that are possible outgroups. PCI, the major size axis, accounts for 67.5% ofthe variation in all taxa examined and PC2, a major shape axis, accounts for 10.3%. PCI ranges fi-om 0.0-6.0 standard deviation units, and PC2 ranges from - 0.5-1.5 standard deviation units.
25 All Taxa Calathaemon Freshwater .
p. antrom. Marine \ SUttkbt^- cn ^ J^aemo^ .-^P^hilta P- P^gio j^acrobrachinm Ua.
P. intermedins P. vulgaris
Leander
PCI (67.5%)
Figure 3.3. Centroids and one standard deviation about the centroids as determined by principal component analysis for all taxa studied. Major freshwater species of Palaemonetes from North America mainly fall into a cluster above the line of separation (—) between fi-esh and salt water species. No salt water species (including brackish water species) fall within the freshwater cluster. The bold elUpses represent Macrobrachium, one ofthe possible outgroups, and P. antrorum an aberrant subterranean species. PCI, the major size axis, accounts for 67.5% ofthe variation examined, and PC2, a major shape axis, accounts for 10.3%. PCI ranges from 0.0-6.0 standard deviation units, and PC2 ranges fi-om - 0.5-1.5 standard deviation units.
26 corroborated by the univariate measurements of distance characters (Appendix). This analysis also suggests a separation ofthe marine from the freshwater species, with the exception of P. paludosus and P. suttkusi which are intermediate between the marine species Macrobrachium and the marine species P. hiltoni and P. intermedins. An important fmdmg is that in no circumstance were any marine or brackish water taxa grouped among the freshwater groups. A possible interpretation is that the origin of this subset (freshwater Palaemonetes) ofthe Palaemonetidae had its origin in sah water and that P. paludosus and P. suttkusi retained some ofthe marine ancestral characters. The problem whh this interpretation is that size pla\'s such a significant role in the separation of these taxa, and separation on size alone could account for this pattern. The univariate measurements of characters (Appendix) indicate that P. paludosus and P. suttkusi are relativel}' large members of this group. PC2 accounted for 10.3% of residual variation and PC3 accounted for 6.5%. PC2 and PC3 were major shape axes with potentially some unparthioned residual size variation. A plot of PC3 versus PC2 (Fig. 3.4) indicated that Leander was separated from all other taxa on the major shape axes, and that P. antrorum was discriminated from all other congeners of Palaemonetes. The centroid analyis of PC3 versus PC2 offered no additional interpretive value (Fig. 3.5). Vector correlations ofthe first two major PC axis of discrimination (Fig. 3.6) indicated that no single character discriminated among the groups but that the characters form a uniform cluster which is heavily size loaded. Vector correlations ofthe major shape axes (PC2 and PC3) indicated that a rosette of characters contribute to the shape discrimination of Leander and other taxa (Fig. 3.7). When examining the first two axes of discrimmation of categorical groups, usmg discriminant function analysis (DFA), three initial clusters arose (Fig. 3.8): (1) P. antrorum and Calathaemon; (2) Leander; (3) all other taxa. Interestingh. P. antrorum and Calathaemon were collected from the same geological formation; an
27 All Taxa
p. antrorum Calathaemon
Leander
CL
P. intermedins
PC2(10.3%)
Figure 3.4 Convex hulls and centroids indicating discrimination due to the two major shape axes as determined by principal component analysis for all taxa studied. PC2 accounted for 10.3% ofthe variation due to shape, while PC3 accounted for 6.5%. PC2 is scaled from -0.5-1.5 standard deviation units, and PC3 is scaled from -1.0-1.5 standard deviantion units .
28 All Taxa p. antrorum o Calathaemon
Leander CO m cu
P. intermedius
PC2(10.3%)
Figure 3.5. Centroids and one standard deviation about the centroids, due to the two major shape axes, as determined by principal component analysis, for all taxa studied. PC2 accounted for 10.3% ofthe variation due to shape, while PC3 accounted for 6.5%. PC2 is scaled from -0.5-1.5 standard deviation units, and PC3 is scaled from -1.0-1.5 standard deviation units.
29 u cu
c o
fc o
Correlation with PCI
Figure 3.6. Vector plot showing character correlations. Arrows indicate the directionality of characters. In this plot all characters contribute relatively equal amounts to the character separation (see Figs. 3.2, 3.3). The length ofthe arrows represent the magnitude of contribution to group separation. PCI and PC2 axes range from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number.
30 m CJ 0.
O CJ
Correlation with PC2
Figure 3.7. Vector plot showing character correlations ofthe major shape axes. A rosette of characters contribute to the shape discrimination of all taxa (see Figs. 3.4, 3.5). Both the PC2 and PC3 axes range from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number.
31 All Taxa
bu Q
p. antrorum Leander
Calathaemon DFl (37.5%)
Figure 3.8. Plot ofthe major shape axis (DF2) versus the major size axis (DFl), as determined by discriminant function analysis for all taxa studied. DFl (size) accounts for 37.5% and DF2 (shape) accounts for 24.4% ofthe variation amoung taxa. The range of DFl is -2.0-2.0 standard deviation units, and DF2 ranges from -2.5-1.5 standard deviation units.
32 artesian well on the campus of Southwest Texas State University in San Marcos, Hayes Co., Texas and from Ezelle's Cave, San Marcos, respectively. Examination ofthe DFA centroid plot offers additional discrimination (Fig. 3.9). The marine and brackish water species {Macrobrachium, Palaemon, P. hiltoni, P. intermedius, P. pugio, P. vulgaris) and the freshwaterspecie s P. paludosus form one group while the majority of freshwater species cluster together. A plot of DF3 versus DF2 further discriminates Calathaemon from P. antrorum (Fig. 3.10). Centroid plots for the same relationship gave no fiirther resolution and are not shown here. Vector analysis resuhed in rosette formations for DF2 versus DFl (Fig. 3.11) and DF3 versus DF 2 (Fig 3.12) with no single character discriminating among groups. The character evaluation using Rao's V, indicated that ofthe 30 final characters used in the analysis the maximum number of discriminatory characters was not additive, but that the maximum discrimination power per character occurs at approximately five (Fig 3.13). As the number of characters are added to the data set, the rate of accumulated information content does not increase (Fig. 3.14). The characters which were discriminatory in the analysis (for all taxa evaluated simultaneously) are given in Table 3.1. The size-free discriminant analysis ofthe first two major axes of discrimination (Fig. 3.15) and the centroid analyses (Fig. 3.16) were qualitatively similar in information content to the analyses using size-uncorrected discriminant function analysis (Fig. 3.9). The conclusions reached were the same: the marine and brackish water species {Macrobrachium, Palaemon, P. hiltoni, P. intermedius, P. pugio, P. vulgaris) and the freshwaterspecie s P. paludosus form one group, while the majority of freshwaterspecie s cluster together. The subterranean species P. antrorum and Calathaemon group together and exhibit certain qualitative characteristics similar to the marine species (i.e., antennal morphology; marine species have the unfused portion ofthe antennae longer than the fused segments).
33 All Taxa
,'P. pugio Palaemon
intermedius P paludosui7 9 acrobrachium CN Figure 3.9. Centroids, and one standard deviation about the centroids, as determined by discriminant function analysis for all taxa studied. DFl (size) accounts for 37.5% and DF2 (shape) accounts for 24.4% ofthe variation amoung taxa. The range of DFl is -2.0- 2.0 standard deviation units, and DF2 ranges from -2.5-1.5 standard deviation units. 34 All Taxa-^ Freshwater \ Leander \ P. antrorum 00 m b Q Subterranean ^ Calathaemon Marin• e N \ DF2 (24.4%) Figure 3.10. Plot ofthe major shape axes, as determined by discriminant function analysis for all taxa studied. DF2 (primary shape axis) accounts for 24.4% and DF3 (secondary shape axis) accounts for 8.6% ofthe variation amoung taxa. The range of DF2 is -2.0-1.0 standard deviation units, and DF3 ranges from -2.0-2.0 standard deviation units. 35 Correlation with DFl Figure 3.11. Vector plot showing character correlations of DF2 (shape axis) and DFl (size axis). Plot indicates that a rosette of characters contribute to the shape discrimination of all taxa (Figs. 3.8, 3.9). Numbers represent the major characters that discriminated taxa. DFl ranges fi-om-1.0-1. 0 standard deviation units and DF2 ranges from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 36 tu Q c o 1 V u. O Correlation with DF2 Figure 3.12. Vector plot showing character correlations contributing to character variation resulting in shape discrimination as determined by discriminant function analysis, for all taxa studied. Arrows show major characters contributing to the separation of taxa (Fig. 3.10). Both axes represent shape components, with DF2 ranging from - 1.0-1.0 standard deviation units, and DF3 ranging from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 37 yTX All Taxa 4.5 .2 4.0 03 > a. 3.5 o »: 3.0 > E 2.5 3 2.0 1 1 1 1 1 10 15 20 25 Number of variables included Figure 3.13. Plot ofthe cumulative Rao's V values for all taxa studied. The maximum discrimination power per character occurs with the fifdi character. 38 Vi O ed E 3 0 10 15 20 Number of variables included. Figure 3.14. Cumulative Rao's V values for all taxa studied. The rate of accumulated mformation content per character decreases with the addition of characters. 39 Table 3.1. Rao's V hierarchy for the 26 characters used in the analyses of all taxa. 1. Dactyl Leg 11 2 Corpus Leg II 3. Telson M x M 4. Corpus Leg 1 5. 3rd Segment Antenna Peduncle 6. Merus Leg IV 7. Telson L x L 8. 2nd Segment Antenna Peduncle 9. Corpus Leg V 10. Ischium Leg IV 11. 3rd Maxili Propodus/Dactyl 12. Carapace Length 13. Corpus Leg II 14. Dactyl Leg I 15. Propodus Leg V 16. Dactyl Leg IV 17. Ischium Leg 1 18. 3rd Maxili Carpus 19. Telson, Base to middle spines 20. Telson, Middle spines to lower spines 21. Ischium Leg V 22. Propodus Leg IV 23. Merus Leg V 24. Merus Leg 1 25. Propodus Leg I 26. Ischium Leg II 40 All Taxa Calathaemon. Marine ' 4 Leander P. antrorum Subterranean t J 00 1/3 Freshwater SFl (38.0%) Figure 3.15. Plot ofthe secondary axis of discrimination (SF2) and the primary axis of discrimination (SFl), as determined by size- fi-ee discriminant function analysis for all taxa studied. SFl accoimts for 38.0% ofthe variation, and SF2 accounts for 24.8%. SFl ranges fi-om -2.0-2.0 standard deviation units, and SF2 ranges from -1.5-2.5 standard deviation units. 41 All Taxa Calathaemon Marine /' ^ 0 Leander p. antrorum Subterranean 00 CO Freshwater SFl (38.0%) Figure 3.16. Centroids and one standard deviation about the centroids as determined by size-firee discriminant function analysis for all taxa studied. SFl accounts for 38.0% ofthe variation, and SF2 accounts for 24.8%. SFl ranges from -2.0-2.0 standard deviation units, and SF2 ranges from -1.5-2.5 standard deviation units. 42 A size-fi-ee cluster phenogram created by using unweighted pair group method using arithmetic averages (UPGMA) ofthe D^ unrooted (Fig. 3 17), suggested three hypotheses. (1) The marine taxa (plus P. paludosus, fi-eshwater) form a cluster. (2) All ofthe remaining fi-eshwater taxa cluster together, (3) A size- fi-ee cluster phenogram created using unweighted pair group method using arithmetic averages the fi-eshwater cluster includes the subterranean species {Calathaemon, P. antrorum and P. cummingi). A neighbor-joining unrooted cluster phenogram (Fig. 3.18) resulted in a grouping that included all ofthe fi-eshwater species, with the exception of P. paludosus. P. paludosus clustered with populations of marine and brackish water origin. The subterranean populations {Calathaemon, P. antrorum and P. cummingi) form a subcluster among the fi-eshwater taxa, indicating a possible convergence in morphological form due to shared habitat constraints or the release of them. This suggests the hypothesis that the marine taxa are ancestral to the fi-eshwater taxa and the subterranean taxa are highly derived fi-eshwater forms. P. paludosus (fi-eshwater in habitat) groups with the marine forms indicating that it may represent a possible second invasion of fi-eshwater,bu t retaining many ofthe ancestral marine characteristics. If this is correct, the genus Palaemonetes is paraphyletic. A neighbor-joining phenogram, rooted at the logical outgroup as determined by the PCA analysis (i.e., Leander) (Fig. 3.19) resulted in the positioning of Palaemon and Macrobrachium well within the Palaemonetes phenogram and between the fi-eshwater and marine clusters. These two taxa are considered to be brackish water inhabitants. These findings may indicate that Palaemon and Macrobrachium may be large marine relatives that are closely allied to Palaemonetes and are convergent to one another due to the shared habitat or shared ancestry. This is consistent with the hypothesis of Pereira (1989) that Palaemonetes is a sister taxon to Macrobrachium. 43 I to u N o c OX) o o o SB u u o U O a es « o a o" •H o p jtf a S o t S o. w > o" « 2 t 3 t a Z o u a = 3 CC > es o a a 5 crj 00 to •^ •2 -2 •c <«5 "a So 2es 2 S !^ •SI?? Ji o 2 O ^ 2 ^ 2 1 1 o O ft > ii; -^ -i c: -:«: .Ac -S S: -is -^ OD C a, c^ a, ci, a. ti.c, :i,a,a,c c:,a,c. ^ c. ^ U 0. o uu^ T O c .2 4—• O o •5 c B to 5 o o c 8 S 00 o o 44 CO ^ c -a 1^ '-5 3 2 Vi ed e« OD o Vi OL, *-H Vi C .2 o SS B o CO U a ex u< OD a O 4> •^ .o c O o. D< o O O u o OD tK. U a C •a o > o a > o .«2 <3 JC 00 -ic *S 2 c 1 "I ~ -c a> -is o •T3 ci. ? 2! S I ~ f\ C o Si == 2 CL l-cl 1 ft.; ^ D J ^;; J ^ o6 R,i V 3 OD 45 B 03 C P u 03 Vi s: r --3 w O O >-i 3 p •Nk "55 5 X ca o OD U c a o o CQ o { O U o o u Vi o U a •E ed Vi a > a _o > •<—> .2" •2 -2 -2 (D Mb 5 5 2 -^ I «0 I •I a ti^ ft 1 2 ^ I L I I I p J5 Ui ^ a s! OD _1_L JL O I ^ c OD *c :2. Q, »- 3 J C (SO o p -J ^ OD 3 -53 o ON rn 3 OD 46 Comparison of Species within the Genus Palaemonetes A PCA was preformed on the genus Palaemonetes, by gender (Fig. 3.20) in order to determine whether there was within-genus character variation that could be attributed to sexual dimorphism. The convex hulls ofthe PCA, by gender, exhibited a significant amount of overlap, with the centroids relatively close together. The centroids on PCI differed by approximately one standard deviation (SD) unit in the full space, and on PC2 by no measurable difference. This indicates that there is little evidence for sexual dimorphism in the size component and no sexual dimorphism on the shape component of this sample of organisms. Since the centroids and convex hulls of this plot did not indicate any sexual dimorphism in character space, the genders were pooled for further analysis. A MANOVA was attempted, to establish an F statistic for gender differences among Palaemonetes, but the individual sample sizes ofthe taxa were not large enough to contribute to a meaningful comparison. The PCA plot indicated that PCI accounted for 70.8% ofthe variation in the characters (Fig. 3.21). The loadings of PCI were all positive, indicating that PCI is a general size vector. Species of Palaemonetes jfrom North America can be separated into three groups based on distance measurements. The first discrimination is represented by P. antrorum, the aberrant subterranean species which Holthuis (1952) placed in the subgenus Alaocaris. The freshwater species {P. cummingi, P. hobbsi, P. kadiakensis, P. lindsayi, P. mexicanus, P. suttkusi, and P. texanus) form the third group (group A). A marine group (group B) consists ofthe species, {P. vulgaris, P. intermedius, P. pugio, and P. hiltoni) and P. paludosus. The centroid analysis further clarified the central tendencies ofthe groups (Fig. 3.22). PC2 accounted for 8.4% ofthe character variation species. Shape appeared to be the largest contributing factor to separation within the fi-eshwater and marine groups. A plot of PC3 on PC2 (Fig. 3.23) maintained the three groups but failed to contribute additional information of separation within the groups. A character 47 00 u PCI (70.8%) Figure 3.20. Convex hulls and centroids indicating the discrimination among males and females, for species of Palaemonetes from North America. PCI, the major size axis, accounts for 70.8% ofthe variation in all taxa examined, and PC2, a major shape axis, accounts for 8.4%. PCI ranges fi-om 0.0-6.0 standard deviation units, and PC2 ranges fi-om - 0.4-1.6 standard deviation units. 48 Palaemonetes p. antrorum A 00 Freshwater u ON PCI (70.8%) Figure 3.21. Relationship between the first two axes of discrimination as determined by principal component analysis; convex hulls with centroids for populations of Palaemonetes from North America. The freshwater group (Group A) consists of Calathaemon, P. antrorum, P. cummingi, P. hobbsi, P. kadiakensis (six populations), P. lindsayi, and P. texanus. The marine group (Group B) consists of Leander, Macrobrachium, Palaemon, Palaemonetes hiltoni, P. intermedius, P. pugio, P. vulgaris plus P. paludosus and P. suttkusi which are freshwater inhabitants. The dotted line indicates a separation ofthe basically fresh and salt water taxa. PCI, the major size axis, accounts for 70.8% ofthe variation and PC2, a major shape axis, accounts for 8.4%. PCI ranges from 0.0-6.0 standard deviation units, and PC2 ranges from - 0.5-1.5 standard deviation units. 49 Palaemonetes p. antrorum P.cummingi Freshwater p. suttsnki P. paludosus 'P. vulgaris P. pugio P. intermedius Marine PCI (70.8%) Figure 3.22. Centroids and one standard deviation about the centroids as determined by principal component analysis for species of Palaemonetes from North America. PCI, the major size axis, accounts for 70.8% ofthe variation, and PC2, a major shape axis, accounts for 8.4%. PCI ranges from 0.0- 6.0 standard deviation units, and PC2 ranges from - 0.5-1.5 standard deviation units. 50 Palaemonetes. p. antrorum m U OH Marine Freshwater PC2 (8.4%) Figure 3.23. Convex hulls and centroids indicating the discrimination among species of Palaemonetes from North America, due to the two major shape axes. PC2, the major shape axis, accounts for 8.4% ofthe variation examined, and PC2, a major shape axis, accounts for 6.3%. PC2 ranges from -0.5-1.5 standard deviation units, and PC3 ranges from - 0.8- 0.6 standard deviation units. 51 correlations vector plot suggests that all the characters contributed relatively equally to the separation ofthe groups (Fig. 3.24). Discriminant ftinction analysis (Fig. 3.25) suggests support for the division of the populations of Palaemonetes into two groups: (1) a marme cluster (plus P. paludosus); (2) the freshwater groups, and P. antrorum as an outlier. Complete discrimination is expressed on both the major size axis and the major shape axis. Centroid analysis provides no more information concerning the groups and is not shown here. DFl accounts for 41.8% ofthe variation while DF2 is responsible for 24.3%. The vector plot of DF2 versus DFl (Fig. 3.26) suggests that the major characters of discrimination form a rosette by which the groups separate. A second discriminant function analysis, DF3 versus DF2 (Fig. 3.27), was completed to determme any further separation within the groups. The freshwater species remain m a tight cluster with no discrimination within the group. However, within the marine group the convex hulls are more resolved. Palaemonetes vulgaris clearly separates from the other marine species. Palaemonetes hiltoni also separates from other members ofthe group. The remaining species only marginally separate. In this DFA, DF3 accounts for 9.9% ofthe variation. An mterestmg issue is that P. antrorum clusters vnXh the marme group. Based on the poshion of P. antrorum and P. paludosus m the neighbor-joining phenogram (Fig. 3.19), it suggests that these two freshwater species have retained similar marine ancestral shape characters. The importance of this is, that it offers additional inference that the freshwater Palaemonetes is not monophyletic. A plot ofthe centroids and one standard deviation (Fig. 3.28), on DF2 and DF3 axes, suggests that the individuals within the marine group do mdeed separate with no overlap of characters at one standard deviation. The vector plot mdicates the characters that separate the different species of Palaemonetes in the second DFA (Fig. 3.29). 52 eg O Q. SI c o o o Correlation with PCI Figure 3.24. Vector plot showing character correlations contributing to character separation among species of Palaemonetes from North America (Figs. 3.21, 3.22). Arrows show a directional component with all characters contributing relatively equal amounts of variation to the separation of taxa. Both the PCI and PC2 axes range from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 53 Palaemonetes "^ cn W Freshwater CM b p. palndosi// O P. antrorum / Marine DFl (41.8%) Figure 3.25. Plot of die major shape axis (DF2) verus the major size axis (DFl), as determined by discriminant function analysis for all species of Palaemonetes from North America. DFl accounts for 41.8% of die variation within the Nortii American Palaemonetes, while DF2 accounts for 24.3%. DFl ranges from -3.0-2.0 standard deviation units, and DF2 ranges from -2.0-1.5 standard deviation units. 54 Q c o fc o CJ Correlation with DF 1 Figure 3.26. Vector plot showing character correlations among species of Palaemonetes from North America. DF2 (shape axis) and DFl (size axis) indicates that a rosette of characters contribute to the shape discrimination of all taxa (Fig. 3.25). Nimibers represent the major characters that are used in the discrimination of taxa. DFl ranges from -1.0-1.0 standard deviation units, and DF2 ranges from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 55 Palaemonetes p. pugio P. paludosus P. antrom^ P. intermedius Q P. hiltoni Marine Freshwater p. vulgaris DF2 (24.3%) Figure 3.27. Plot ofthe major shape axes, as determined by discriminat function analysis for populations of Palaemonetes from North America. DF2 accounts for 24.3% ofthe variation and DF3 accounts for 9.9%. DF2 ranges from -2.0-1.5 standard deviation units, and DF3 ranges from -3.0- 2.0 standard deviation units. 56 Palaemonetes P. pugio P. palndosns\^_j P. intermedins P. 0antrorum Marine e / Freshwater p. vulgaris / DF2 (24.3%) Figure 3.28. Centroids and one standard deviation about the centroids for the major shape axes, as determined by discriminant analysis, for species of Palaemonetes from North America. DF2 accounts for 24.3% of die variation and DF3 accounts for 9.9%. DF2 ranges from-2.0-1. 5 standard deviation units, and DF3 ranges from-3.0-2. 0 standard deviation units. 57 Palaemonetes 29 Q § 28 ^y^ 30 c o at i J^^^>> 10 orre l ''^^^*r^ CJ 8 Correlation with DF2 Figure 3.29. Vector plot showing character correlations of DF3 (minor shape axis) versus DF2 (major shape axis) among species of Palaemonetes from North America (Figs. 3.27, 3.28). The plot indicates that a rosette of characters contribute to the shape discrimination studied. Numbers represent the major characters that are used in the discrimination of taxa. DFl ranges from -1.0-1.0 standard deviation units and DF2 ranges from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 58 The character evaluation using Rao's V suggests that, ofthe 30 final characters used in the analysis, the maximum number of discriminatory characters was not additive and the maximum discrimination power per character occurs v^th the addition ofthe fifth character (Fig 3.30). As the numbers of characters were added to the data set, the rate of accumulated information content did not increase (Fig. 3.31). The characters that were discriminatory in the analysis (for all taxa evaluated simultaneously) are given in Table 3.2. Size was removed from the data set and a size-free DFA was completed. The size-free DFA suggests a consistent separation ofthe same groups as above. However, the SFl vector accounts for only 42.3% ofthe variation (Fig. 3.32). The second and third axes ofthe size-free DFA, do not suggest further separation within either group (Fig. 3.33). A size-free discriminant function analysis cluster phenogram based on the D was created using unweighted pair group method using arithmetic averages (UPGMA). This unrooted phenogram (Fig. 3.34) suggests marine taxa (plus P. paludosus, freshwater) cluster together. The remaining freshwater taxa cluster together including the subterranean species {Calathaemon, P. antrorum and P. cummingi). A result of a neighbor-joining unrooted cluster phenogram (Fig. 3.35) is the grouping of all freshwaterspecie s except P. paludosus. P. paludosus clusters with populations of marine and brackish water origin. The subterranean populations {P. antrorum and P. cummingi) form a subcluster among the freshwatertaxa . The marine taxa clustered together. Comparison of Discrete Characters As stated in the introduction, several discrete characters have been used in an attempt to resolve the systematic relationships within the North American species of the genus Palaemonetes. The egg size and number are consistent in all species examined, that is the freshwater forms have larger but fewer eggs than are found in 59 10 15 20 Number of variables included Figure 3.30. Plot ofthe cumulative Rao's V values for all species of Palaemonetes from North America showing the maximum discrimination power per character occurs at approximately the fifth character measured. 60 0 10 15 20 25 Number of variables included Figure 3.31. Cumulative Rao's V values for all species of Palaemonetes fromNort h America showing that the rate of accumulated information content per character decreases with the addition of characters after the addition ofthe fifth character. 61 Table 3.2. Rao's V hierarchy the 30 characters used in the analyses of Palaemonetes. 1. Telson L x T. 2. Propodus Leg II 3. Telson M x M 4. Corpus Leg II 5. Corpus Leg V 6. Telson B X M 7. Corpus Leg II 8. Telson L x L 9. Merus Leg IV 10. Propodus Leg IV 11. Ischium Leg V 12. Rostrum length 13. Carapace Length 14. Corpus Leg 1 15. 3rd Maxili Propodus/Dactyl 16. Merus Leg V 17. Dactyl Leg II 18. Ischium Leg IV 19. Propodus Leg 1 20. Dactyl Leg IV 21. Ischium Leg II 22. 3rd Maxili Carpus 23. Ischium Leg 1 24. Telson M x L 25. 3rd Segment Antenna Peduncle 26. 2nd Segment Antenna Peduncle 27. Corpus Leg IV 28. Dactyl Leg 1 29. Mems Leg 1 30. Propodus Leg V 62 SF2(24.8%) Palaemonetes Figure 3.32.Plotofthsecondsize-frediscriminatoryaxis the variation,andSF2accountsfor24.8%.lrangefrom North AmericanPalaemonetes.SFlaccountsfor42.3%of -2.0-3.0 standarddeviationimitsanSF2rangefrom-2.0- (SF2) vs.thefirstsize-frediscriminatoryaxis(SFlfor 1.5 standarddeviationunits. Marine ^N/V \W) ^^^ \ J^x^N/^NP-ontromm^^ ^1^ 4:j]^^^) Freshwater SFl (42.3%) 63 ^-•-• - Palaemonetes P. pugio I p. paludosu. P. antrorum C/3 Marine Freshwater p. vulgaris SF2 (24.8%) Figure 3.33. Plot ofthe third size-free discriminatory axis (SF3) vs. the second size-free discriminatory axis (SFl) for the North American Palaemonetes. SF2 accounts for 24.8% of the variation, and SF3 accounts for 10.2%. SF2 ranges from -3.0-2.0 standard deviation units, and SF3 ranges from -2.0 to 1.0 standard deviation units. 64 es Vi S B o 3 o O o lO d o c lO a> o 0) lO ^ "2 d c i_g> O o 5 i> o 8 E "2 is 3 o 'X 15 -J CQ <0 2 .2 -2 CO Co" 52 CO 55 c- - ^ ,- c: C C •5 CO CO c c =; (b 0} CO 0} m fl> CO .2 .^ -c: •to -25 .2 C' D «_o c .CO .CO S m i .2 SJi•^S ^ -9 fl5 "O "O 5 "5 "6 -S "5 .2 p 5 *V ^U 0^ ^W ^W C MJ (D 3 O O CD 5 a CO %: a r^ CO .C ^ .^ £ ^ .^ :% <« a; a: Q: CL QJIKI; 0:0:0.0.0:0.0.0.0.0. o: 0: uy^ tn ^ o c a Ui C8 u - \ . - o 22 o c CO cd o o « Vi ch I N O O m 65 ^1 a. 3i ^1 c o CO Vi ex s: Vi fc C 1^ o U S ed 0^ hd o SS 00 O o O .A d U c •53 u Cu > _o o .5 '^ 00 a s u ee o .5 = o c J r- 2 > ca ee .5 _l I -^ Si. 1 to 3 00 66 the marine species. While P. kadiakensis remains the exception, the freshwater species exhibit a reduced number of larval stages than the marme. The fused segments ofthe antennular flagellum also serve as a discrete character to separate the marine and freshwater species. While P. antrorum and P. holthuisi are exceptions to this character, it may be they have retained ancestral characters that may benefit them in a tactile environment. The segmentation on the antennal scale in the form I zoea presents some questions. Palaemonetes kadiakensis, a freshwater species, exhibits a segmented antennal scale as do the marme species. Also, P. paludosus, another freshwater species, has an unsegmented antennul scale in the form I zoea. In all ofthe analyses performed in this study, P. paludosus always groups with the marine species even though the segmentation ofthe form I zoea is consistent with the freshwater species. The grouping ofthe freshwater species may be due to sraiilarities m sediment load ofthe environments inhabited by this group of species. The movable exopod spination as a possible systematic character proposed by Strenth (1976) does have some support in that P. hobbsi and P. mexicanus group together and share the same movable expod condition. Palaemonetes lindsayi and P. suttkusi exhibit a lack of exopod spines and group together. However, the subterranean form P. antrorum, which shares the same exopod condition as do P. lindsayi and P. suttkusi, is grouped with the other subterranean species. The species P. texanus appears mtermediate within populations of P. kadiakensis; however, P. texanus exhibits the same condition ofthe movable exopod spine as does the P. hobbsi and P. mexicanus group. Larger sample sizes may help to resolve the relationships based on the condition ofthe movable exopod spine. Two additional characters were discussed in the mtroduction, rostrum dentition and appendix masculinae, as being considered as possible characters to resolve systematic questions of this group. As was pointed out, the variation within each species is large and overlaps among the different species. 67 Comparison within Species of P. kadiakensis Six populations of P. kadiakensis fromwithi n the state of Texas were examined. In order to determine the effects of sexual dimorphism, a PCA was performed, by gender (Fig. 3.36). Examination ofthe convex hulls for PC2 versus PCI shows males and females overlap significantly suggesting a lack of sexual dimorphism within these populations of P. kadiakensis. The within-species comparison contained a larger sample than the comparison across all taxa or within the genus Palaemonetes. Because ofthe significant samples size ofthe pooled P. kadiakensis populations (N~60), a MANOVA was successfully applied to examine the degree of statistical difference between males and females in multivariate space. The F statistic indicate males do not significantly discriminate from females in the full character space (F 30,6= 0.43; p=0.94). Because this was found, data were pooled across gender for fiirther analysis. The PCA for the subset of P. kadiakensis was performed and PCI accounted for 78.1% ofthe variation (Fig. 3.37) within the characters. This suggests that most ofthe separation within these populations is due to size. Only 3.2% ofthe variation is accounted for by PC2. The centroid and one standard deviation plot for PCI versus PC2 (Fig. 3.38) more clearly defines the separation ofthe populations. With the exception of a minor overlap with the highly variable Tom Green Co. population, the Hidalgo Co. population is separate fromth e remaining five populations. The centroids of four ofthe six populations lay along the major size axis (PCI) and cluster together. It is interesting to note size plays an important role in the separation ofthe Hidalgo Co. and Val Verde Co. populations, which both are part ofthe Rio Grande River drainage system. The Val Verde Co. population exhibits marginal separation fromth e other populations with only a small amount of overlap between the ellipse for the Menard and the Val Verde Co. populations. There appears to be a large shape component within the variation for 68 / \ P, kadiakensis / \ ' \ / \ / \ \ / \ / \ . \ y-^ / \ /^ ' " rr\ \ C/ Males • ^Females C2( : \ cu \ ' \ ^^^^ \ \ A'^'''''^'^ \ I ^ \ \ •"" —. __ \ \ ~~ ~~ ^^ .^ \ ^•^"---^^^ PCI (78.1%) Figure 3.36. Convex hulls and centroids indicating the discrimination among males and females, for populations of Palaemonetes kadiakensis. No sexual dimorphism was evident (FJQ^ = 0.43; p = 0.94). PCI, the major size axis, accounts for 78.1% ofthe variation in all taxa examined and PC2, a major shape axis, accounts for 3.2%. PCI ranges from 1.0-6.0 standard deviation units, and PC2 ranges from - 0.6-0.8 standard deviation units. 69 P. kadiakensis / \ Hidalgo Co. Val Verde Co. / • (I^^ ^ Menard Co. > \Tom Green Co. Burleson Co. Llano Co. ^"^^^^^ \ PCI (78.1%) Figure 3.37. Relationship between the first two axes of discrimination as determined by principal component analysis; convex hulls with centroids for populations of Palaemonetes kadiakensis studied. PCI, the major size axis, accounts for 78.1% ofthe variation in all taxa examined and PC2, a major shape axis, accounts for 3.2%. PCI ranges from -1.0-6 standard deviation units, and PC2 ranges from -0.6-0.6 standard deviation units. 70 p. kadiakensis Hidalgo Co. Val Verde Co. CN Menard Co. CN u cu Burleson Co Tom Green Co. PCI (78.1%) Figure 3.38. Centroids and one standard deviation about the centroids as determined by principal component analysis for populations of Palaemonetes kadiakensis studied. PCI, the major size axis, accounts for 78.1% ofthe variation in all taxa examined and PC2, a major shape axis, accounts for 3.2%. PCI ranges from 1.0-6.0 standard deviation units, and PC2 ranges from -0.6-0.6 standard deviation units. 71 the Tom Green Co. population. The vector plot for PC2 versus PCI (Fig. 3.39) suggests that all characters contribute relatively equal in the separation of these six populations. There is very little shape separation in PC3 versus PC2 (Fig. 3.40), each axis accounts for only about 3% ofthe variation. However, there is marginal separation ofthe Burieson Co. population. All other populations cluster with little separation of the centroids ofthe convex hulls. The centroid and one standard deviation plot for PC3 versus PC2 (Fig. 3.41) suggest the Burieson Co. population does separate from the remaining populations with no overlap. The vector plot (Fig. 3.42) suggests the most significant character in the separation ofthe Burieson Co. population is the ischium of leg one. DFl accounts for 86.4% ofthe variation, while DF2 accounts for 10.3%. All populations discriminate in DF2 versus DFl (Fig. 3.43). The DF3 versus DF2 (Fig. 3.44) plot also discriminate the six populations. However, the relationship ofthe populations from Hidalgo Co. and Menard Co. is seen to be farther apart, and the Burleson Co. population becomes more intermediate between the other two populations. However, this may be artifact of reduced sample size. The centroids and one standard deviation plot for DF3 versus DF32 do not provide any additional informative information and are not shown here. The character correlation vector plot of DF 2 versus DFl (Fig. 3.45) suggests the characters used in the separation of populations is directional, with no one character having a significant effect on the discrimination of this group of Palaemonetes. The vector plot for DF3 versus DF2 (Fig. 3.46) also suggests no one character has a significant effect. The size-free discriminant fimction strongly differentiates the six populations. SFl accounts for 67.4% ofthe variation while SF2 is responsible for 25.2% (Fig. 3.47). There is no additional information contributed by the centroid and the one standard deviation plot, so this diagram is not figured here. The vector plot (Fig. 3.48) is represented as a rosette with no single character having a 72 Correlation with PCI Figure 3.39. Vector plot showing character correlations. Arrows show a directional component with all characters contributing relatively equal amounts of variation to the separation of populations of Palaemonetes kadiakensis (Figs. 3.37, 3.38). The PCI ranges from -1.0-1.0 standard deviation units, and PC2 ranges from -1.0-1.0 standard deviation units. See Table 2.2 for die list of characters by character number. 73 ' P. kadiakensis Burleson Co./ • >v./ y/^ / \Tom Green Co. X ^ fl^^^^ X 1 ^ 1 / X m# ^ *» ^^^^^^\^^^^ 1 '' 1 MB ^ \ "^^..^^^ Hidalgo Co. I / 1 / 1 ' \ • 1 j.f t / \ • 1 ^ / 1 /I / >/^ 1 \ '^\. i / / w\ / 1\ • 1 Menard Co. ^ t - - " "" /^ 1 , ^ Val Verde Co. Llano Co. j PC2(3.2%) Figure 3.40. Convex hulls and centroids indicating the discrimination among all taxa due to the two major shape axes for Palaemonetes kadiakensis populations studied. PC2, the major size axis, accounts for 3.2% ofthe variation in all taxa examined and PC3, a major shape axis, accounts for 3.0%. PC2 ranges from -0.6-0.6 standard deviation units, and PC3 ranges from -1.0 to -0.2 standard deviation units. 74 P. kadiakensis Burleson Co Menard Co. cn cn OH Tom Green Co Hidalgo Co. Val Verde Co. Llano Co. PC2 (3.2%) Figure 3.41. Centroids and one standard deviation about the centroids due to the two major shape axes, as determined by principal component analysis, for populations of Palaemonetes kadiakensis studied. PC2, the major shape axis, accounts for 3.2% ofthe variation in all taxa examined, and PC3, a minor shape axis, accounts for 3.0%. PC2 ranges from -0.6-0.6 standard deviation units, and PC3 ranges from -1.0 to -0.2 standard deviation units. 75 P. kadiakensis Ischium Leg I 29 cn UQ_ ith i 13 \ J Jf 25 ^ c o "^ 2 ^^y% -re ! 16 ^^10 o U Correlation with PC2 Figure 3.42. Vector plot showing character correlations. Arrows show a directional component with all characters contributing relatively equal amounts of variation to the separation of populations of Palaemonetes kadiakensis (Figs. 3.40, 3.41). PC2 ranges from -1.0-1.0 standard deviation units, and PC3 ranges from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 76 P. kadiakensis \ Val Verde Co. Llano Co. ^ Menard Co. cn b O Burleson Co. 0 CN 0 PU Hidalgo Co. O Tom Green Co. DFl (86.4%) Figure 3.43. Plot ofthe major size axis vs. the major shape axis, as determined by discriminant function analysis for populations of Palaemonetes kadiakensis studied. DFl accounts for 86.4% ofthe variation while DF2 accounts for 10.3%. DFl ranges from -1.5-1.5 standard deviation units, and DF2 ranges from -2.0-2.0 standard deviation units. 77 Hidalgo Co /A P' kadiakensis ^ Val Verde Co. Burleson Co. p CN Menard Co. m • Q Tom Green Co. Llano Co. V DF2(10.3%) Figure 3.44. Plot ofthe major shape axes, as determined by discriminant function analysis for populations of Palaemonetes kadiakensis studied. DFl accounts for 10.3% of the variation, while DF2 accounts for 2.0%. DFl ranges from -2.0-2.0 standard deviation units, and DF2 ranges from -2.5- 1.5 standard deviation units. 78 Correlation with DFl Figure 3.45. Vector plot showing character correlations for populations of Palaemonetes kadiakensis. Arrows show a directional component vsdth all characters contributing relatively equal amounts of variation to the separation among populations (Fig. 3.43). The DF2 ranges from -1.0-1.0 standard deviation units, and DF3 ranges from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 79 Correlation with DF2 Figure 3.46. Vector plot showing character correlations for populations of Palaemonetes kadiakensis. Arrows show a directional component with all characters contributing relatively equal amounts of variation to the separation among populations (Fig. 3.44). The DF2 ranges from -1.0- 1.0 standard deviation units, and DF3 ranges from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 80 p. kadiakensis \y Tom Green Co. Burleson Co. Hidalgo Co. CN CN O CN I 7 Menard Co. tJU 17 V Llano Co. Val Verde Co. ^ SFl (67.4%) Figure 3.47. Plot ofthe major discriminatory axis (SFl) and the secondary discriminatory axis (SF2), as determined by size-free discriminant function analysis for the genus Palaemonetes kadiakensis. SFl accounts for 67.4% ofthe variation, and SF2 accounts for l'S.1%. SFl ranges from -2.0-2.0 standard deviation units, and SF2 ranges from -2.0-2.0 standard deviation units. 81 Correlation with SFl Figure 3.48. Vector plot showing character correlations. Arrows show a directional component with all characters contributing relatively equal amounts of variation to the separation among populations of Palaemonetes kadiakensis (Fig. 3. 47). The SF2 ranges from -1.0-1.0 standard deviation units, and SF3 ranges from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 82 significant effect on the discrimination of these populations. The convex hulls for SF3 versus SF2 (Fig. 3.49) also discriminate all ofthe populations ofthe genus of Palaemonetes from North America. The Burleson Co. population moves into a more central position withm the group of populations. And the Tom Green Co. and the Val Verde populations appear to shift the relationship between these two groups with Burleson Co. and the Menard Co. populations falling between the Tom Green and the Val Verde populations. The centroid and one standard deviation plot did not add any informative information. Rao's V (Fig. 3.50) was used to examine the value added by each additional character. Because ofthe close genetic relationship ofthe populations, more characters may be needed to see subtle within-species discrimination. The slope of this relationship begms an ascend greatly at 25 characters, and the information content at the maximum number of characters (i.e., 30) is likely due to accumulated noise and not accumulated character value. A major finding ofthe within-species comparison is the amount of discrunination evident among the six P. kadiakensis populations from Texas. The separation pattern follows the paradigm suggested for maximum evolution of closely related taxa. That is, if the populations are in restrictive drainage systems isolated for long periods ofthe populations have the potential to become independently dissociated from the marine ancestral forms m various and autonomous manners. The plot for the cumulative Rao's V (Fig 3.51) indicated that as characters were added to the Rao's V analysis the rate of information per character added decreased after the addition ofthe tenth character. The characters which were discriminatory in the analysis (for all taxa evaluated simultaneously) are given in Table 3.3. The unrooted UPGMA phenogram representing the size-free Mahalanobis distance analysis (Fig. 3.52) indicates a possible clustering ofthe Tom Green and the Val Verde populations and a second cluster containing the remaming populations. 83 P, kadiakensis Val Verde Co. Hidalgo Co. Burleson Co. to m tu Menard Co. Tom Green Co. V Llano Co. SF2 (25.2%) Figure 3.49. Centroids and convex hulls, as determined by size-free discriminant analysis for populations of Palaemonetes kadiakensis studied. SF2 accounts for 25.2% ofthe variation, and SF3 accounts for 4.5%. SF2 ranges from -2.0-2.0 standard deviation units, and SF3 ranges from -2.5-1.5 standard deviation units. 84 0 10 15 20 Number of variables included Figure 3.50. Plot ofthe cumulative Rao's V values for all populaitons of Palaemonetes kadiakensis studied showing the maximum discrimination power per character. See text for detailed discussion. 85 60 . 50 . CO I 40 > i2 3 30 . E 3 CJ 20 . 10 . 10 15 20 Number of variables included Figure 3.51. Cumulative Rao's V values for populations of Palaemonetes kadiakensis studied showing that the rate of accumulated information content per character decreases with the addition of characters. 86 Table 3.3. Rao's V hierarchy the 30 characters used in the analyses of Palaemonetes kadiakensis. 1. Corpus Leg V 2. Ischium Leg II 3. Corpus Leg 1 4. Ischium Leg 1 5. Telson B X M 5. Propodus Leg II 7. Dactyl Leg 1 8. Merus Leg V 9. 3rd Maxili Carpus 10. 3rd Maxili Propdus/Dactyl 11. Carapace Length 12. Rostrum length 13. Propodus Leg 1 14. 3rd Segment Antenna Peduncle 15. Corpus Leg II 16. Ischium Leg IV 17. Propodus Leg V 18. Merus Leg 1 19. Ischium Leg V 20. Dactyl Leg II 21. 2nd Segment Antenna Peduncle 22. Telson L x T 23. Telson M x M 24. Dactyl Leg IV 25. Corpus Leg II 26. Telson M x L 27. Telson L x L 28. Propodus Leg IV 29. Merus Leg IV 30. Corpus Leg IV 87 o o a 6 6 uo •p U 6 O a o •p QC u o > ^ o e. s a > a o w O CJ CN C cd .4—• Q o o so O O 00 88 A similar pattern of clusters is supported by the unrooted neighbor joining phenogram (Fig. 3.53). However, there appears to be a shift among the populations ofthe Burleson Co., Hidalgo Co., Llano Co. and Menard Co. cluster. 89 5J O 3 C O > > O U o a o u o o es o 90 CHAPTER IV CONCLUSIONS Several consistent patterns emerge from this work: (1) Leander appears to be a valid outgroup for the North American species of Palaemonetes. This possible ancestral connection along whh the fact that there are extant marine species of this genus indicates that Palaemonetes likely arose in the marine environment, rather than arising in freshwater and re-invading the marine environment as hypothesized by Perefra (1989); (2) Calathaemon appears to be a highly derived species of Palaemonetes. In every analysis, Calathaemon appears within the genus Palaemonetes rather than separated from the group. This research provides several hypotheses for the systematic relationships of Palaemonetes from North America. Systematic reevaluation ofthe taxa indicate that there is little evidence to support the hypothesis by Bruce (1993) that P. holthuisi should be elevated to a new genus Calathaemon. It should remain in the genus Palaemonetes and the name of P. holthuisi restored to its origmal designation as stated m the rules of taxonomic nomenclature. The elevation of P. antrorum to the subgenus Alaocaris is not supported by this research and I feel P. antrorum should remain within the genus Palaemonetes. Palaemonetes antrorum appears to be a highly derived species within the genus Palaemonetes; however, it appears to be an intermediate sister taxon among other species in the genus. The group containing P. antrorum, P. cummingi, and Calathaemon require ftirther study to understand whether this group is indeed a monophyletic group. All three species live in a subterranean environment, so it is highly likely that this group exhibhs convergent evolution rather than a close systematic relationship. The data in this study supports P. paludosus to be more closely related to the marine species than to the other freshwater groups, and fiirther suggests the freshwater species have arisen on at least two different occasions. However, P. paludosus exhibits the freshwater conditions associated with this group, egg size and 91 number, reduced larval stages, and the non-segmentation ofthe antennular scale in the form I zoea suggests P. paludosus is more closely associated with the other freshwater species of Palaemonetes from North America A possible explanation for the grouping of P. paludosus with the marine group may be parallelism from the retention of ancestral characters or convergent evolution due to similarities in sediment loads within their environments. The data suggests the remaining freshwater species of Palaemonetes form a monophyletic group representing a single invasion into freshwateran d this group has undergone radiation since that invasion. Therefore, further study is needed to resolve the hypothesis by Sollaud (1923) and Strenth (1976) proposing North American freshwater Palaemonetes form a monophyletic group. Dobkin's (1965) proposal that P. kadiakensis is a recent immigrant into the freshwater environment is not supported. Palaemonetes kadiakensis appears to be ancestral to other freshwaterspecie s of this genus. The segmented antennule ofthe first zoea of P. kadiakensis would appear to be due to convergent evolution with the marine forms. The extended larval development stages exhibited by P. kadiakensis may in fact be a resuh of selection because ofthe wide spread and variable environmental conditions found across the range of this species. The marine and freshwater taxa represent two divergent groups, probably separated at some ancient time and then radiating within their respective habitats. However, P. paludosus appears to be more closely related with the marine species than to other freshwatergroups . The neighbor-joining phenogram indicated that the marine species are probably ancestral to the freshwater species. This would suggest that the freshwaterspecie s have arisen twice and therefore do not form a cohesive group, as proposed by Sollaud (1923) and Strenth (1976). However, the remaining freshwater species of Palaemonetes do appear to form a paraphyletic group however. Subterranean species fiirtherdiverge d in a convergent manner to exploit that extreme habitat. 92 Several biogeographic patterns also emerged. The three subterranean species, P. antrorum, P. cummingi, and the genus Calathaemon, are consistently grouped together. The two species fromth e eastern slope ofthe Sierra Madre Oriental Mountains in Mexico, P. mexicanus and P. hobbsi repeatedly appear to be sister taxa. While P. suttkusi and P. lindsqyi, both fromth e west ofthe Sierra Oriental Mountains form a natural group as well. The populations of P. kadiakensis appear to be broken into two groups depending on their geographic range. One group of P. kadiakensis, includes populations that are east and south of a line following the Balcones Escarpment in central Texas. The second group of populations of P. kadiakensis are found north and west ofthe line. Palaemonetes texarms, fromHaye s Co., Texas appears to separate the north and western populations from the eastern and southern populations. Within the out groups included in this study, Palaemon and Macrobrachium appear to be closely related; however, Leander appears to more closely aligned with the marine group of Palaemonetes than to either Palaemon ox Macrobrachium. When Leander is included in the analysis, it always appears closely related to P. vulgaris. 93 LITERATURE CTFED Benedict, J. E. 1896. Prehminary descriptions of a new genus and three new species of crustaceans from an artesian well at San Marcos, Texas. Proceedings of the United States Natural History 18:615-617. Bookstein, F., B. Chemoff, R. Elder, J. Humphries, G. Smith, and R. Strauss. 1985. Morphometries in Evolutionary Biology. Special Publication 15, The Academy of Natural Sciences of Philadelphia. 277 pp. . 1989. Size and shape: A comment on semantics. Systematic Zoology 38:173-180. Broad, A. C. 1957. Larval development of Palaemonetes pugio Holthuis. The Biological Bulletin 112:144-161. Broad, A. C. and J. H. Hubschman. 1962. A comparison of larvae and larval development of species of eastern U. S. Palaemonetes with special reference to the development of Palaemonetes intermedius Holthuis. The American Zoologist 2.394-395. . 1963. The larval development of Palaemonetes kadiakensis M. J. Rathbun in the laboratory. Transactions ofthe American Microscopical Society 82:185-197. Bruce, A. J. 1993. Leptoplaemon gagudjui gen. nov., a new freshwaterpalaemoni d shrimp from Arnhem land, and a re-evaluation of Palaemonetes holthuisi Strenth, with the designation of a new genus, Calathaemon. Hydrobiologia 257:73-94. Chace F. A., Jr. 1954. Two new subterranean shrimps (Decapoda:Caridea) from Florida and the West Indies, with a revised key to the American species. Journal ofthe Washington Academy of Sciences 44:318-324. Colless, D.H. 1970. The phenogram as an estimate of phylogeny. Systematic Zoology 19:352-362. 94 Collins, J. T. 1993. Systematic implications of allozyme variation in freshwater species of Palaemonetes (Crustacea, Decapoda) from Texas and Mexico. Master's Thesis, Angelo State University. 44 pp. Creaser, E. P. 1932. The decapod crustaceans of Wisconsin. Wisconsin Academy of Sciences, Arts, and Letters 21.32\-22%. Creaser, E. P., and A. I. Ortenburger. 1933. The decapod crustaceans of Oklahoma. Publications ofthe University of Oklahoma Biological Survey 3:13-47. Digby, P. G. N., and R. A. Kempton. 1987. Multivariate Analysis of Ecological Communities. Chapman & Hall, London. Pp. 206. Dobkin, S. 1963. TheX^rvdX deyeXo^ment of Palaemonetes paludosus {Gibhes, 1850) (Decapoda, Palaemonidae), reared in the laboratory. Crustaceana 6:41-61. . 1965. Abbreviated larval development in caridean shrimps with the description ofthe larval stages of some species from south Florida. PhD. Dissertation, University of Miami. Pp. 228. . 1971. The larval development of Pa/ae/wowe/e5Cw/w/w/wg7 Chace, 1954 (Decapoda, Palaemonidae), reared in the laboratory. Crustaceana 20:285- 297. Farris, J. S. 1972. Estimating phylogenetic trees fromdistanc e matrices. The American Naturalist 106:645-668. Felsenstein, J. 1984. Distance methods for inferring phylogenies: A justification. Evolution 3%. \6-24. Fleming, L. E. 1969. Use ofthe male external genitalic details as taxonomic characters in some species of Palaemonetes (Decapoda, Palaemonidae). Proceedings ofthe Biological Society of Washington 82:443-452. Garcia, D. K. 1991. Population structure and patterns of speciation in the genus Palaemonetes. Doctoral dissertation, Texas A and M, College Station, Texas. 131 pp. Garcia, D. K., and S. K. Davis. 1994. Evidence for a mosaic hybrid zone in the grass shrimp Palaemonetes kadiakensis (Palaemonidae) as revealed by multiple genetic markers. Evolution 48:376-391. 95 Gibbs, L. R 1848. Catalogue ofthe fauna of South Carolina. In: Tuomey, M., Report on the Geology on the Fauna of South Carolina. Appendix, I-xxiv. 1850. On the carcinological collections ofthe cabinets of Natural History ofthe United States: with an enumeration ofthe species contained therein, and description of new species. Proceedings ofthe third meeting ofthe American Association for the Advancement of Science 3:165-201. Heller, C. 1869. Zur naheren kenntiss der in den sussen gewassem de sudlichen Europa vorkommenden meerescustaceen. Zeitschrift fuer Wissenschaftliche Zoologie 19:156-162. Hobbs, H. H., Jr., H. H. Hobbs, m, and M. A. Danial 1977. A review ofthe troglobitic decapod crustaceans ofthe Americas. Smithsonian Contributions to Zoology 244:l-\S3. Hobbs, H. H., in, and J. P. Jass. 1988. The crayfish and shrimps of Wisconsin. Milwaukee Public Museum, Milwaukee, Wisconsin. Pg. 114-121. Holthuis, L. B. 1949. Note on the species of Palaemonetes (Crustacea: Decapoda) found in the United States of America. Proceedings ofthe Koninklijke Nederlandse Akademie van Wetenschcppen 52:87-95. . 1950. Subfamily Palaemoninae. Part I in the Palaemonidae collected by the Siboga and Snellius Expeditions, with remarks on other species. Part X in the Decapoda ofthe Siboga Expedition. Monograph 59c? in Siboga Expeditie. 268 pp. . 1952. The subfamily Palaemoninae (Crustacea:Decapoda:Natantia) of the Americas, n. Alan Hancock Foundation Occasional Papers 12:1-396. Houck, M. A. 1992. Morphological variation in an ectoparasite: partitioning ecological and evolutionary influences. In: J. T. Sorensen, and R. Foottit (eds). Ordinations in the Study of Morphology, Evolution, and Systematics of Insects: Applications and Quantitative Genetic Rationales. Elsevier, Amsterdam. Pg. 277-308. Houck, M. A., J. A. Gauthier, and R. E. Strauss. 1990. Allometric scahng in the earhest fossil bird, Archaeopteryx lithographica. Science 247:195-198. 96 Houck, M. A. and B. M. OConnor. 1998. Morphological variation in Hemisarcoptes (Acari: Hemisarcoptidae): Application of muhivariate morphometric analyses. Annals ofthe Entomological Society of America 91: 335-349. Hubschman, J. H. 1974. The larval development of Palaemonetes intermedius Holthuis, 1949 (Decapoda, Palaemonidae) reared in the laboratory. Crustaceana 26:89-103. Hutcheson, J. H., J. H. Oliver, Jr., M. A. Houck, and R. E. Strauss. 1995. Muhivariate morphometric discrimination of nymphal and aduh forms ofthe blacklegged tick, Ixodes scapularis (Acari: Ixodidae), a principal vector of the agent of lyme disease in eastern North America. Journal of Medical Entomology 32:827-842. JoHcoeur, P. 1963. The multivariate generalization ofthe allometry equation. Biometrics 19:497-499. Kluge, A. G., and J. S. Farris. 1969. Quantitative phyletics and the evoultion of annurans. Systematic Zoology 18:1-32. Lowe, B. T., and A. J. Provenzano, Jr. 1990. Survival and reproduction of Palaemonetes paludosus (Gibbs, 1850) (Decapoda: Palaemonidae) in saline water. Journal of Crustacean Biology 10:639-647. McCutchen, H. L. 1983. Morphological variation of appendicies masculinae in North American species of fresYiWdXer Palaemonetes (Crustacea, Decapoda). Master's Thesis, Angelo State University. 30 pp. Meehean, O. L. 1936. Notes on the freshwatershrim p Palaemonetespaludosua (Gibbes). Transactions of the American Microscopical Society 55:433 -441. Neilsen, L. A. and J. B. Reynolds. 1977. Population characteristics of a freshwater shrimp, Palaemonetes kadiakensis Rathbun Transactions ofthe Missouri Academy of Science 10:44-57. Nelson, D. H. 1982. Thermal tolerance and preference ofthe freshwatershrim p Palaemonetes kadiakensis. Journal of Thermal Biology 7:183-187. Pereira, G. A. 1989. The cladistics, taxonomy, biogeography, and evolutionary history ofthe shrimp family Palaemonidae (Crustacea, Decapoda, Caridea). Doctoral Dissertation, University of Maryland. 418 pp. 97 Rathbun, M. J. 1902. Descriptions of new decapod crustaceans for the west coast of North America. Proceedings ofthe United States National Museum 24:885- 905. Say, T. 1918. An account ofthe crustaceans ofthe United States. Journal of the Natural Academy of Sciences of Philadelphia 1:57-458. Smalley, A. E. 1964. A new Palaemonetes from Mexico (Decapoda: Palaemonidae). Crustaceana 6:229-232. Sollaud, E. 1923. Le developement larvaire des "Palaemoninae. " Bulletin biologique de la France, et de la Belgique 57:509-603. Strauss, R. E. 1985. Evolutionary allometry and variation in body form in the South American catfish genus Corydoras (Callichthyidae). Systematic Zoology 34: 381-396. Strenth, N. E. 1976. A review ofthe systematics and zoogeography ofthe freshwater species of Palaemonetes Heller from North America (Crustacea: Decapoda). Smithsonian Contributions to Zoology 228:1-27. . 1991. The significance ofthe larval development of Palaemonetes antrorum Benedict (Crustacea, Decapoda). On the origin of subterranean palaemonids. Stygologia 6:149-153. . 1994. A new species of Palaemonetes (Crustacea: Decapoda: Palaemonidae) from northeastern Mexico. Proceedings ofthe Biological Society of Washington 107:291 -295. Strenth, N. E, and G. Longley. 1990. Reproductive patterns of the subterranean shrimp Palaemonetes antrorum Benedict (Decapoda, Palaemonidae) from central Texas. Stygologia 5:221-224. Ulenhuth, E. 1921. Observations on the distributions and habits ofthe blind Texan cave salamander, Typhiomoige rathbuni. Biological Bulletin 40:73-104. Vandel, A. 1965. Biospeology. Pergamon Press, New York. 524 pp. Villalobos, A. and H. H. Hobbs, Jr. 1974. Three new crustaceans from La Media Luna, San Luis Potosi, Mexico. Smithsonian Contributions to Zoology 174:1-18. 98 Webb, D. J. 1980. Electrophoretic systematics and population genetics ofthe genus Palaemonetes (Decapoda, Caridea) in Texas. Master's Thesis, University of Houston. 98 pp. Wiley, E.O. 1981. Phylogenetics; The Theory and Practice of Phylogentic Systematics. John Wiley and Sons, New York. 439 pp Worth, S. G. 1908. Freshwater shrimp, a natural fish food Bulletin of the Bureau of Fisheries 28:853-858. Wright, 1954. The interpretation of multivariate systems. In: O. Kompthome, T. A. Bancroft, J. W. Gowen, and J. L. Lush, (eds.). Statistics and Mathematics in Biology. Iowa State University Press, Ames. Pg. 11-33. Xuan, N. V. 1997. On a new species of prawn, Palaemonetes camranhi n. sp. from brackish water of south Viet Nam (Decapoda:Caridea). Crustaceana 70:85- 94. 99 APPENDIX 100 Si^f2l5S23tSS£!^0"!? oor^coi«^o>o>^ U r^a>CMror^a>'<4-oooo'^ooor^o>inr<~co(Ooo(Ooooocoa>'«-o ooootocococoo-^-iOooo-^ioi^^oor^-cjioaooooo 5 incMinT-T-h-0(OT-o>cM(or^oooT-^cama>-<-cM(MCNjoco 2 ^ CNiiocv4^r^(pcvjcoroa>oq(pa>T-oqc\j'^a|oqco-^'^a>r^^u) «^ I—I c o h- CMCMfM"«-T--.-T--r-'*-(MCO-r-COCNJ"«-CN4CN4CNjT-"«-T-T-T-^C4'r- o cococo •«-• C4-1 o >> H.3 3 ooa>ooo>oooor^ooa>(Oi^<0(0(0<0(Doo(0(0(or^ro(0(ot*^(0 2 N 3 o V) M (0 Vi (0 CO W) M 3 3 3 3 3 3 3 3 3 >s >s >» >» >s >» >s >s 3 C C C C C C C C (Q CO CO CO CO CO CO CO a. .2 .2 .2 .2 .2 .2 .2 .2 .2 CO (cS (0 (0 CD (S (0 (Q (0 (O (A M (/> (0 (O (0 (0 o a>o)o>o>o>o>D>o>o> X X X X X X X X X T3 "O T3 "O •0 •a •0 •a 333333333 0) Q> (D 0) (D C c c c c c c c I cu Q.Q.Q.Q.Q.Q.Q.CLQ. a> 0) 0) 0) u c 0) ^ ^ CO _0)jgja)_© _ o> 0) a> a> CD CO CO CO a PPP.2.2J2.2PPEE— E — — — ~EEE — — — — E — i U.U.U_SSSSlJ.Ll-Ll.U.SU-SSSSLJ-li.U-2SS2lJ_S o U ^r\iroir>rnis.«iri,-0<*>"**'>tf>r^«)0>OCO"*mCDr«-000>0 U 3 rs.r«-i^r^r^is.rooooo h-CMCMCMCNjcsjcMCMcor^r^r-h-r^r^i^oo O "o > T-rsico^»ocor>- 101 c>|ioo^^c3>(N^r^r^^coco^h«.r«~oooiocgcooco^cDCDco 00 £ eo^ooocococDOJ^oi^cgocoo^ooco^'^com^i^^^ lOT-iocoT-^^ocO'T-oocOT-in^r^csj^^csjT-coooco -- CO rooooococsioococomtncM^ r^'»-cocor^^ <3>CO-«-(\4'^lOOOr^CDCOCN4inO>CDCDCOCO'«-^CO^OOOO^OO"* «/)ocoooojo>oO'«-co«ncooo <3>eococMr^oo-«-CMr^cocooocM CMcorg-T-cgojcgcMCMcacgcMcgcococgcococgcMOicoojcMCMCM (Ococor^cDr^-cocococor^cDcor^r^r^coh^i^r^h-r^miocou) a o (/}(/)(0(/>(/)(/)(/)(/)(/>(0 3 3 3 3 c c C C >s>sC0CDCOCOC0C0CDCOCDCcocDcoco(/)(0(/)(n(/)(/)(0(0(/0 ) 3 QCOOOOOOOOOOO 3333333333(^(/>(0(0 O. T7 .J a^^ a^^ a^^ aai^ a^^ •••* a^^ a^^ a^^ a^^ o vy V) X X X X X X X X X x^^^^i^i^^^:S-H-y"y Pu :i:iEEEEEEEEEE 3333333333 O O O O X C/5 EEE— EE— E — — E — — — EEEEE — — — E — — E ll_U.LJ.SlJ.LJ-SU-2SlJLS2SLJ-ll-U-li.LLSSSlJ.:E2U- •o a> JC •»-CMco^»ocDh-oo 102 oo>cDco»ncoin(3)0>co^«o ^ ir^£^<'*'-^*>^o*^f^ pr^^oooocM'v-T-cocorgcooocococoinT-ooo^cocoooin coco'^ocoooocoooo>coininoo(7>oo^cgh«-r»oor<-incMcgoo S JS corgin(ococMT-c3>-^(3>cD^< CO eg o o -«- ^ in m 00 o> o> T-O CO CO oh-'«tcoooinincococoT- h- CO CM eg ino m 1^ "^ o "* o 1^ CO o T-cj>CMin"*co^^cocoh~ 00 (M o CM •«- ^ o o> CO in CM 1^C M r^ 00 i>»-cginco^-r-co ^ h- 00 T- o> r^ o 00 COo 1^ r>^ ^ •«- <3) COo o o> CO a> p p CO p (M o ^ CO in CO u moocoom^T-cMcocM^coooco^ CMcod^odr^dcocdodocf •»-T-T-CM-«-CMfOCgCMCMCMCM CM CM CO •«-•«- CMCMCMCMCMCM'^CO^CO'^'^COin^COOOOOOOOOOO inincocococO'«-oooo>oo>a>o>ooooo>^T-'«-a>T-^ocMCM _C5 •-' c: o *S (/)(/>(0(/>(/>(/)(/)CO(/>(AEEEEEEEEEE 3 m m m m m m -c -c •= -c -= -= -c -c •= -c E E E S E S E S 5 5 ex o .Q^.Q.Q.Q.Q«WWCOCOCOCO«OBeo5§555§^i^§ OH 0000003333333333CCC"c'c'c'c"c'c'c J:=.C:XI.C.C.C>>>>>>>>>>COCOCOCOCOCOCOCOCOCD X J2 .2 J2 i* ii .2 .2 _a> _© _« _a) ii .2 c/3 CO CD CO CD OB CO 0) c 0) CO CO CO CO CO CD ro 103 h*CMoocoooinoo^o>oocoo>eOT-T-coo)r^ino>cMO>coincMh- ooco-<-coinco<7>co^cocMCDin-«cM^T-c»ocOT-pr-^coo•CM-«-ooO'«-oocgoooooooin-«-'^co^coincor^t:;^'«"r-J>^ copco^'^^r*.in^cMoocO'»-o>oo>pppppi^PP'^ Le g iliu m •T-^T-^T-^f-T-^^^l-^^-r^T-^^T-^d^dddd^-r^^'-'-^ CO o HCHO ocM^cor^oor^cooocMcomcoco-'-ooocMt-cocDpiniOT-co v ^ocoincgcor^cgT-or^ooincooocococO"*cj>oo^^cpin^ U JS comcococMooocoococMCMh-i^r-.oo'^ooo^'^'-^t^SJO CO copcginr*-coininin<»^ini^«incop'^copp«cO'«-;^p CX u CO O r- T- T- d^'r-^T-^T-^r^r^r-^r^o>i**cdt*-^h--o6d'*-cMoo>oo u. c CO Ja > CMin^<3>i^'* o>CMT-inT-o>incoo> CMr^h-cgcoo-^CMmmococo-^oocM . ^inocomoocooooo ^oo H O cM-»-oooh-ot-CMor^coi^t^r*-»*»-t*-t*-i^i**c3>r*-r»-i^h*r^ -^ CJ a> CX o xxiiiixxi>>>>mmoQ 3 ccccccccccWWOT«rocDCDcoco«co5roro_CO _C O CX o CO CO CO CD CO CO CO CD CO CD^S^^HD^^^^^TOHD^^'-O^ OH (Da> X 2 .2 ^ ^ Jg 're'cDCOCOCOCDCDCDCOCOCOCOCO^. n CO CO CO EEEEEEEEEEEEE-2E-2i2i2i2EE-32-2EECD CO E T3 CX 104 108 7 006 3 389 8 152 8 117 0 029 5 078 4 285 9 545 0 444 6 855 4 826 1 404 6 665 7 223 9 215 6 440 0 3 .911 8 .476 1 ,100 7 ,091 8 ,367 5 .599 2 .102 6 .306 6 .416 5 2 o CO O HCHO CO 103 7 870 8 788 6 373 9 236 8 920 4 952 1 520 0 911 3 688 3 293 7 654 0 .325 1 .427 5 .669 1 U .621 5 .143 8 .765 3 .978 5 411 9 ,232 0 .223 6 .434 2 .846 3 .627 4 .144 4 CO CO 1^ CO '^ 00 oo ^ O) CM ^ CO r*-00 Oi h- 00 ^ Oi o oo 00 o CO a. c ^ ^ ^ Ca r a JS 127 8 409 4 536 8 525 8 866 2 660 9 877 7 055 4 378 7 794 3 263 4 647 4 242 3 259 5 271 0 677 2 .154 6 2 U .784 7 ,011 6 ,735 8 .365 3 .715 9 .515 1 .629 7 .870 3 .714 1 -s G ^ in ^ oo •* in CO in oo r^ a> in CO h* in CO h* 1^ h- CO "* 1^ C3> 00 00 00 V e^ 069 6 106 1 525 8 548 0 829 6 560 8 750 7 521 0 705 4 383 1 705 6 215 2 942 1 248 3 498 5 642 2 673 2 406 7 ,777 0 ,170 2 415 9 .416 1 .420 2 .885 6 .205 3 .159 7 "* CM CO CO "* 00 in CM CM m -* CO 00 in r^ 1^ 00 m CO Oi in CO CO I 3 T— CM CM CM CM CM CM CM CO IM CM CM CM CM CM CM CM CM CM CM CO CM CM CM CM CM CM CO CM CO CM CO CO CO CM CO CO ^1 h- 1^ 00 CO r^ CO CO 00 r*.r» - oo r^ 00 r- r^ 00 oo r^ h«. r^ CO 00 o> 00 Oi 00 JS a> CX H D c o o o o o tn (0 CO Ui m CQ m S S S S S S S S S S _l _i —I —1 ••-o» K h- 1- h- 1- —I —t CO 3 CX o kadia k - kadia k - kadia k - kadia k - kadia k - kadia k - kadia k - kadia k - kadia k - kadia k - kadia k - kadia k - kadia k - kadia k - kadia k - kadia k - kadia k - kadia k - kadia k - kadia k - kadia k - paludos i paludos i PU paludos i paludos i paludos i X {/oJ Femal e Mal e Mal e Femal e Mal e Mal e Mal e Femal e Femal e Femal e Mal e Mal e Mal e Femal e Femal e Mal e Femal e Mal e Femal e Mal e Femal e Femal e Femal e Mal e Mal e Mal e l>N CJ 39 9 •o 86 9 3 JS 65 8 65 9 66 0 66 1 66 3 66 4 66 5 66 6 3 65 6 65 7 68 4 65 4 65 5 68 5 68 7 65 3 69 6 70 0 61 7 65 2 69 7 69 9 Ou 61 5 61 6 C > U o m CO oo o CM CO "* in CO rx- oo Oi o CM CO in CO 00 a> o o o o ^ CM CM CM CM CM CM ^ G o o CM CM CM CO CJ < a CJ o Xi «> CO CX H c« 105 <5h*COI^"^^<3>OOCMCOOOOOC»r^CMT-COCMCMOOOOO>COinr^p incoo>r^cMcocM^coi^'^^0)'«-o>coo>ooininin'^oo^inin co^ininoor^cocooo-«-ino>cor^'^cococMOinoor^'«-cO'«-o> 3 oboJKpflqo>coo)d>^f^r>lcMT^Sb>d>pc6pcMT-;inpp 1 1^ T-dd-r-^ddddd-r-^^dddd'T-^^^cMCM^^ CO ^M ogcocMcoincMoo^inco^CM-«--t-cMoocMinoooo^ooinoo a> oor^ooinoiT-cocMcocoini^^oocoo^mrMcocno^in "* T- o JS PT-odioico^cdcjVfindiof^cMS^oocMcocoo)cococoo CM CM CO p CO p r*-copT-coppp"^"^pcMppcMpppcMh«; CX u CO c ^ CO CO CM cd r^cMi^odcdcdcda>odcdcdcdr^a>r<-a>indinco ^— ^— CM •r- T— Ca r ^cOT-co^inT-cooor^cocM^r^cocoo>cocoh-^0)0>r*-cMi^ ^comoococo^h-h-^coT-comcM^cocDOCMoococom"*'*- a JS cO"^0'«-r^i^cor>-coinh«-cj>cO"*oor^^c3>i>»-cMcocor»-coi^co ^ ppph-;'r-CMOCOinT-OOCMCOp^CMO>pp'«-;pppCMp^ 2 cdd^^l^"l^lncd^»-'lncd•^ln•^lnln•^'^"•^lncdlnlnT-^co•»-^co Le n Ros t r^cocoi^inc3)OCMC3>in'«-o>"^cor^ooc3>incMor^r*cooT-r>. ^c3>oinc3>i^inoO'«-0"*cj>oi^oooocMcor^CMr^r^ococo^ a ^cRcoocow-^co-T-CRinh-T-cMooincooooooor-^^^CMf^ ru m ppph«-CM^ppr>--ppp^^CMCMpp'<-;"^^pp'«-:pp -•i 3 CO uUl T-din^T-^in'od^cdh^cococoodobcO'^CMinooinod'^d'^op O CM CM CM CM CM CM CO ^ CO CO Pd O coco^coeO'^mcocO'^'^^inco^cococococococoininin'^ j-i 4> CJ O H ooooooooooooh^oor^^ooh^r^^oooooooooooooo ±1 « ^ CX a> CX UiUiUiUiUiUiUiUi o Ui Ui Ui Ui CO .3 .3 3 3 3 3 3 ^ 'M 3 3 3 3 3 -o -O "O -a -O -O -D -O t_ i_ 3 O O O O O c c C C C C C C — — — — — — — — —00 S 2 CX •o-o-a-o-oEPPPFPEtccccccccccci O U o CD CD OH ^.2.2.2.2©©a)a)«a5©©2222222£2J^J^ Q. Q. Q. Q. Q. .E .E .E .E .E .S .2 .E S Z S Z S S S Z Z S S X 106 O h- CM 1*^ 00 CO CO •- QO E oo>h-c3>coincocor«-^-incooocococo ^ -^ CM in '^ppp'r-'r-;CMpininr>-T-o>co^in CO CO o o T^ d d d CO O CO ocpco^-incooocMCMcooo o o o o a> 00 -^ CO coo>T^cDO>r»-T-oocooooo o o o o ^ ^ o> CO O JS oomoocgmS-in r^ m o CM T- "o o CM ^ C3> •«-ooi>^0"*cocoin"*^ o o o o o o ^ CO 00 CO COO'^-OO^-OT-O'^T-W^W^^^^^V.^Wo o o o o ^o 00 "^ in h*- po^co^incMomococot-cocooo C3> r«- CD CM 2 ^ ^pCMp^pT-;pp'r-pcMinr»-inp r^ •r- •r^ p odincdr^i^r^r^dt-^dcMCMcdcdinin cd CM d d Ro s ocoor^CMcocomoin o in ^ t- ininc3>c7>CMOCMOoor^ 00 00 1^ CM a COCOQOO'^I^OO-'trCMOO CO O CJ> 00 ru m p ^ ^ p •<-j 3 ppph'-pp^'^cop CO cU>l ^cdi*-^cj>dcMdcdcdd ai d d d O CMt-1-T-CMCMCMCMCOCO COCOCOCOCOCOCOCO'd-COCMOCO^CO"* > fc o Teet h Lowe r go CJ •o C/J CO ON a>a>a>ooooooa>T-oin<<^cocococo O CX c o c c c c c c c c c c •=.•=.•=.•=. oooooooooo.-..-.2'2*2*2' E E E E E E E E E E 52 -JQ E .E .E = 1 (ua}a>a)a>a> X 4) •o •«- CM „,, CO "* CO a> JS 3 i-cMco-^incor^oocRoSSc^SScNi C 3u COCOCOCOCOCOCOCOCOTfSjSl^SSj^ tCi3 o O > U r-ooooT-cMco-^mcor^-ooooT-cM ^" c ininincocococococococococoi^h>r^ < a 4> '3 X> a> CO CX H C/5 107 CM'»-r«-oocoo>r>-coocor^ mcOT-mincMoo^T-ocM-^pcocD ooco^ooor^i^h-oino incoincoinoocooocj>incMr>^co5 00 c o>^ooooh<-ocor^cMr^in 4) 5 ^^^ooooo)i-'*-r^o '^COOOOT-OT-COCM'*f^t^f^OOO> cooco^co^incMcoocj>ocor*--r- •«-• "r^ T^ d T-^ -r^ d t-^ T-: T-" CM CM •^•.-•-r-^T^-r-^-r-'^dT-^-r^-r^'.- o>inT->^oincMcocMCMcor^cocococoi^ococo»ncocO'»-r-^ COCOCMCO^OCOOOOOOOT-COinCM^OOCMt^COCOO>OCOCOCM oocoo^ocoooo>^^o>inT-cM^o>^'«-0)in^cooinco l_J -^^ pppppoh>aooinS>a> ^^T-or-moor^r^coooro "O CO ^•"-"•r^dd-r^dddT-'T-^dT-^ •«-^-»-^T-T-0000-«-0 oocooocMcococoh-^'^cooo^^h-coT-cMi^cominT-oor^ 60 S COOCOmcOO^CMh-OO^OJCOCMCOO'^^OOOCOCOOOCOOO ^^coinT-oocoincooT-ooocO" cococj)incocooocooT-ocj>^h«-^T-'^T-inr^oocO'^^oo r*-t^cooincMC3>cocMcococMh-coocoi^oincj>'«-0"^ininin CMCMOcooocMcOT-cocoi^r^^^cj>CM"*^cj>i^ino"*cooin inT-"^i^cj>c3)oooT-c»coo^coocMcoocM^coin^ocooo CO CO CO T-T-^CMCM^COC0^C0C0C0C0"*C0C0CMCMCMC0C0CM t-cor^incMoo CMOincj>i^CMi^^c3>ooi^^CMr>^inh- coincMcoincMinoocMoocMooT-cj>in^CMCM"*incj>oocMCMco^ coooin^>cMincM^CMcoinc3>ooincor^cocoincotf>cMC7>c7>o GO -- ot^ooh-oin"^i^CM^in"<«r^c3>T-c»oooinooco^"*oO'«-in p^pppi^ppi^ppi^ph>:pppppcoininininT-in d-^-r^ddddddd-r^dddddddddddddy^d 42 O coooooT-oo^r-cM^cocoo>r--CMcoococM'»-coooooo)i^T-i^ CO coT-ini^T-f-T-incocomcMOcooooJO^-ocMOCMOinf^oo 00 3 cocMr^cM^cR^h-CMcooocMcoincM^in-T-inoJOO^ininr^ 0) 13 r-T-cj>T-cgco^^^r^h^cor*-^"^cocococMCMT-'«-oooo •»- CM CM "«- 05 O cR^cooh-co^^m-^CMinoocoi^T-i^i^r^cMcooocMi^mo ^^mcRCMOooco-^coo^ococMCMT-incoooT-h-m^coco a> COincOCMCOCO^OOCJ>CMCOOOCOCMCOOOCMr>-CMOOCO-«-T~-«-cOO 3 CM^T-h--^CMppppp^^pr-;pCMp'^p^T-0O0OI^CM G COCOCOT-'cMCMCMCMCMCOinCM^COCMCMCOCOCMCMCMCMx-^CMCvicM o CJ ino>^ooocOT-in^o>cooooooinr*-o>o>ooococj>cor^'*co f-inco^'«-cOT-ininr>*o>CMCMO>CMino>r^cocMin^^^-oocoin (30 CO to-coi^moooT-cMCMr^oo^T-oco^cMcocMOT-ooh-oooin < coinooincor^coooooT-^ooincM'«-coh--«-o>inco^ooo>i^ CM CM CM COCOCMCOCMCMCMCMCMCM 1 22 H 108 coooo>o>oocoo>^-o>inoo>cor^coo>coinh-CMCMco^^O)«n ^o>oo>h«-cooT-o>ooo)incMoooococoinr^^-^-coinopoocR 00 c oocooo^cO'«-a>cO'<-in^a>a>a>T-cMO>coi^ o> r^ ^ o r*- CO coT-moT-cMT-OT-T-ooooT-oco^ CO ^ CO o •»- T- in o O) CO O T- OT-CMOCOlO^COmi^T-CMCO^OCOOOOOCOOOCMOr^CMO 00 ^ •^ocoocoo^T-ooocMo^coinmcoini^cROoo^eooo oocococo^in'^oo>h-CMCM^oococof^ooococoooinr^T-^ i-J ••-• pP'r-p0>a>000000r>^h~C0OOT-CMC0'<-CMI^C7>000>ppp dd-r-ddddddddd-r^y-y-y-r-^r-^y-ddddddd oinc3>inoinr^c3>h>cococoi^^inT-cococo^CMCMoO'Tfin^ COCOCO^"^C»h«-COT-COCMOOCOCO"^r^OOOOOI^r*-COCMOOCM^ 00 S incMcococMinoo"*cj>Tf-<-oooor^h-"*CMincocOT-T-cot-cocM HJ O ppppT-poOh».CJ>I^COCO'^^CJ>OOCOT-;pCMT-;CMpp^ CMCMCMT-^CMCM-^-^T-^T-^^T-^-r^CMCMCMCO-^'cMCO^CMCMCM-r^^CM G Q r^com-^-coT-cooocococM oinT-^o>^cococO'»-T-oor^ooco COC}>OOCOOOOOOI^COCM ^C»CM^OC3>int»>.CO , ^CMr-cooo>»^^ino>cor^r^'»-'«- 2 ^ ppcMcgp^ppT- r^cMflOT-cor^^o-«-^r^T-o>h-T-inh- •^CM^COCOCOCMCMCOCMCMCMCOCOCOCininT-ocMr^C3>^oor»>-'*ppcMpCMO ^ CM ^ CMCOCMCMCMCOCMCMCO coocooincoin^ininT-o>o>coQocoa>a>oooor^coininr^co CMOCOI^OOinCM-«-COCOr«-OOCMC3>-«-CMCJ)OOI*-C3>CO"t^C3>OCO 4> S2 (OOcooinT-T-ooT-r^cM^h-a>ococooinooocor^ooa>in HJ 2 •^flO"*ooooT-r«-co^coh-"*^c3>incooT-inco-«-ocMcO"*CM CMT-^CM-r-^-r^CM-r^-r^CM-r-T-^-r^CMT-^CMCMCOCMCM^CMCNCM-r^T-^CM CM "^ co^r^oo>*-"^"*coincocMi^ C3>CMOOincO"*COCJ>^C3>COh-^CO •^COCDOCOOCMCOOOCOin^CM 00 G CMCOCMCOC3>OCMCMT-ininC3> om-^T-cMinh-cocoococMco-f- l^h-OOCMOO^in-f-OCMCMO "^CMoooom-^i^ocOT-coooflom J ,3 ocoi^ocj>or^cj>CMc»c»^ COCMCMCM^CM-r^T-^CM-r^T-^-r^CviCMCMCMCOCMCMCMCMCMCMT-^-r^CM C^ ^ cocMcooococooocM'«-ocNCMh-r^incMoocor*.^inoooT-CMO T-incor^coh-ocRcocococoincom^ocoooin-^ooo'^ooo 00 —. •*^COOCOOin"*OT-CM^CMCJ>^CMT-Oh-r^'^OOI^T-^CM h^ppi^pi^ppr^pppi^pppppr^pcocoro-t^inp dddddddddddddddd-r^ddddddddd 42 Q •«-o>^'*cooinco^ocoininooino>cooinocM'^cocor^co CO CMCJ>h*CM^inin^in^-«-ooi^ocMinincMco^co-«-coT-^c» coinincj)Oi-co^r»-T-cM to o oocM-<-cMC7>inooinr^coinc3>oor^cMooooincMooooincMoo>co inocococMCMooT-r*-T-cMr>^cocMOT-coo>o^c3>h--«-incoco h-coo>coooincMooo>^'r-T-o>coo>'*ocMincooococOT-oco 3 OOT-CMpppCMpppppT-h*ppT-T-r..ppCMOO>OOCO G "J e- CMCMCOCMCMCMCMT-^CMCMCM-r^CMCMCMCMCO'cMCM-r-^-r-^CMCO-«-^-r^CM C o •OCMCMOO>CMCOT-r>-inOOOO^'^COT-000>00>OCMOOO>0> poocRcocjiooocooor^-oo^CMcor^OT-o^-^oooooom^^ 00 CO « 3 OCMC3>COT-COT-T-CM"*CMCMCOCO"*COOO"^inT-COCOCOCOCMO> < coh*^o>p^ph^"«-i^r^^pi^<>JpppcMppr*-cor^r«*o CMT-'cMT-^^CM-r-T-^cg^^T-^CMT-^CMCMCMT-^CMr-^T-^-r^CMT-^T-^CM i 5:s 109 iu m Le g 1.103 1 1.246 0 1.114 0 1.472 2 1.089 6 1.379 9 1.283 4 0.980 7 1.346 3 1.154 9 1.422 2 1.692 9 1.441 4 1.319 0 3.020 2 1.457 6 1.162 2 1.425 3 1.075 2 1.369 3 1.317 0 1.628 4 2.024 6 1.080 8 1.245 4 2.147 1 <^ »—^ sc h 00 CJ >-. 964 8 839 4 177 9 914 2 912 5 046 8 576 4 796 2 631 8 608 1 521 4 351 9 515 4 425 3 431 5 .704 1 .889 6 .793 1 HJ .758 2 .074 7 .859 9 ,969 9 ,827 8 .628 0 .837 0 .079 0 CM CM CM T3 CO o ''" o o o o o o o o o o ^~ o o o C (N Q 00 CO CJ 3 179 7 747 1 028 7 584 0 241 0 694 7 666 6 .018 1 .071 1 .260 0 .470 1 .265 6 O .705 9 CM.160 1 ,898 3 CM,181 9 ,858 6 CM,182 1 h-,054 0 CO CO CM CO "*,122 8 CO,667 7 ,129 3 .291 1 .466 1 .117 8 .888 8 T3 CX '^'" ^ ^ •* "* "* c o CN cu 00 M CJ 778 3 883 9 823 4 212 4 3 670 3 498 9 613 4 .257 6 .011 9 .744 3 .600 2 H-) ,380 4 ,272 6 ,723 0 ,042 8 ,319 0 ,472 0 ,311 7 ,368 9 ,291 6 ,750 9 .139 5 .236 3 .801 1 .900 7 .953 2 CO CO CM CO CM CO CO ^ ^ CM CM CM '" CM CM CO ^' CM CM '*" ^ ^ CM CM T3 CO G ^ CN u 00 0) CO 547 8 882 8 850 2 375 7 302 4 022 0 978 2 249 7 651 6 225 0 467 8 ,672 7 ,512 1 ,721 0 ,171 8 ,653 6 .248 1 ,626 6 ,709 8 .399 5 .216 0 .343 4 .925 8 .110 3 .316 4 .333 2 2 CM CM CM in CM CO CM CO CM CO "* a> '^ •*" '^ '" '" 2n d 00 qp u a 544 7 612 4 ,126 5 ,968 7 ,593 3 ,138 8 ,076 7 ,794 9 .888 4 ,425 1 .127 0 .201 2 .714 5 .125 7 .368 2 .665 3 .357 4 .555 8 .708 3 .620 3 .462 8 .424 5 .953 7 .232 4 .407 4 .434 6 CM CM '" CM CM CM "«t ^ •- CM CM CM '" CM CM CO ^ ''" '^~ ''"" ''" ''" ''" ''" ''" ''" CN HM schi u 00 541 7 954 4 729 6 763 0 670 3 696 8 a> >» ,701 5 ,795 6 ,788 8 ,755 8 .761 6 ,669 5 .602 5 ,908 7 .795 8 .054 0 .551 7 ,807 4 .842 3 ,928 8 ,625 3 .699 2 ,893 0 ,512 1 ,708 9 ,621 2 HJ o o o o o o o o o o o o o o o ^~ o o o o o o o o o o ••-» CO ^ Q 00 3 1.661 2 1.129 2 1.531 1 1.463 5 1.216 6 1.759 7 1.229 8 1.241 7 1.596 2 1.501 0 1.426 7 1.434 8 0.899 4 1.369 1 1.355 7 1.297 5 2.138 0 0.833 6 1.376 6 1.113 4 0.988 3 1.378 1 0.809 7 0.998 3 +- CX 2.025 9 0.892 6 to O 00 S2 0) 142 6 799 9 828 3 637 3 726 0 270 3 612 6 279 9 0) 3 ,151 1 ,257 8 ,065 1 .237 4 .155 6 .984 1 .815 0 .242 2 .653 8 .444 8 .455 0 .848 7 .970 0 .618 6 .382 7 .693 7 ,311 6 ,246 9 CM CM CM CM CM CM Tt CM CM CM CM ''" ^ CM CM CO o 55 00 CO 194 9 039 9 859 0 965 0 330 8 859 9 389 1 995 5 398 4 323 2 424 6 .361 9 < ,848 0 ,960 7 .896 8 .934 1 .481 9 .164 7 .173 4 .107 0 .595 4 .770 8 .475 8 .285 4 .489 0 ,171 1 -J g CM CM CM CO 1^ CM eg CO 2S no t^O'«-0'»-r^r*.^'^oocoincDcooco^o>'^cooocMooo>ooco 00 c CMCMI^Oh^OOCOCOOOh-OOOCOOOCO^-^T-COOOO^-CMinOh* CMcocomcMinininoooocoT-ocMor^ooo^— mco^-inooco .2.1 i^^(ococoincMco^r^a>T-o>ooooo^a>ococO'<4-'^pp .l-l 'G -r^'r^-r-^T^^^CMT^-^T-'d-^dT-'T^^d-r-dT-^T-^-r^-^T^T^-T^ ^ h>~ CO 00 CO . ooinoincMO>coooo>co^oooin^o>ooinpp T^coin-r-o>ooT-"^inooh-incoooi^^or*-cM^^ 00 _ r- to Oi ^ t^ r^ooc3>ocMOCMr>-C3)r»-ooooooincMi^cD5th* r*- o> T- 00 o> CO CO•^cooi^oor^cooot^coi^cooooppppi ^ •—• o CMCMCMCMCMCMcocMCMcodddddddddd-r^^dddd T3 CO o m ^ CM ^ m r^ CN ^ ooo>o>cOT-^h*o)'^oor^^o coin^o>r^^i^cj>oocooo^o CMh-co-T-cDcococOT-comh-o ooocoinT-oo^cMinr««."^r>-cj> 00 3 ooooooooocor^ooooincMCM ^CMCM-«-CJ>r^0O^-CO^'«-^CM pp^T-p-^p^cocomoo"* HJ O ^o)r^^i^cor^cOT-;ppph-- "^^•^•^^"^cd^^inT-^-r-^T-^ T-^'r-''r-t-:^^-»-'cMCMCMCM^^ G O ooincocMcooooino"*CMcocoT-i^coco^c3)OCMr-ininoo coooo>T-o>ooo>^T-0'^in'^o>ino)incococoinooco^i^oo r»-cO'«-"*incMinT-coin^r^coocoinoinoco CM^i^inc3>^incooocMCMcoc3>ocoocO"*c»^cj>"^^inincM r^cococoo>r»-h*.coincooooi^eMoocj>"*oincococ3)oor^cj>oo <30 ^ ocMoocoT-cj>ooh*cMcoT-T-h-ooincoh-CMininocooh-in coco^inco^i^coincMcococ3>incoininin-^cocMoo'^ppi^ J 2 CMCMCMCMCMCMCMCMCMCO^T-^T-'^'^-r^T-^-r-T-^^CMCMCMCMCM^ CN '^ ^inoo-^coT-co-T-inoo-i-^cM-^o^h-CR^cBcoocMinr^ cocO'«-"«-co^^inincMT-^coin^incMoco^cMocoo>co^ ^ a t*-C0"^l*-.CMC0CMC0C0inOinT-OO00'«-00C0OCMOC0C0^h- CMCM^inco^'*"*incoooocMCJ>ooi^ooo-«-cocj>cj>coinco CMCMCMCMCMCMCOCMCMCMT-CMCMI-CMCM CM -f- CM CM CM CM CM CM *§ 1 c5 ^ CMr>-<»oocMCMr^ocoino^ocoincj>cj>ocMcor^incj>r*-oin inoinh*c3>ocoini^coooT-coinincMr*«-"^r<«.^h-in-«-ooc»cM 00 — ooT-inmcocMCM^-i-CM-^T-cM-^ooooo^t-^h-oo-^oocjJco 0) >^ CM^cocMcoph>;pppppi^ppppppppi^a>cocoin t—I o •r^-r^^^^f-^T-^-r-^T^T-^dddddddddddddddd cocMOOO>oocOT-'^coo)CMor^incoh*r^'^incocooor»-ocM CO OC»r«-^COCO^CJ>OCM'0"CJ>COOOOOO"*CJ>COI^OO"*CO^CO 00 3 ^inooor^ooh>.cocoo)^ooo)coinc3>o>cocooooocMinr>.ooco f-C0CM-«-CMT-;P"^p^ppi-i-;^ppT-ppp^inCMCMO i a CMCMCMCMCMCMCMCMCMCM-r^d'r^'r^'«-^d-r-^T-''r^-r-^-r^T-''r^T-^T-^.,-: 05 O r^CM0>CM0>inO'^^O't-00CM00CMC00>0>l^'»-C00>00CMC0C0 ^•o>CMT-coincM'T-r^incocoh*oocoh*CMooT-^r«-i>^inco^oo a> h~cooocMcoococMocoT-T-coincoo>coco'^r^^inT-cot-T- 3 0>h-000>OOpph-;ph^^PPPpj**;t^_pph-;^T-in'«J-C0 C •T-^T-'T-^CMCMCM^^CMT-^CMCMCMCMCM-r^T-^-r^CMCMCOCOCMCMCM c o U oinooooococMCMCMinh-o^"^cocMCMcocor>«-ooo<»r*-ini^ oooino^h-o^cM^--«-cj)inoo"*r*-oocj>"^^T-cocO"^^in 00 CO CMC3>COO^COO>'r-COinCMinCM'r-COCJ>inCOCJ)COCJ>inoOOCOC3> 4) 3 < OCRT-CMT-COOOCMCRCO^COmCOh-^CO-^COCOCM-^^T-CJ)!^ 5S CMT-CMCMCMCMCMCM^CM CM CM CM CM T- 111 CMinoocMcooocoinoocj>oO"^coT-"^i^^incoinr-incoo>cj>;U coh-r*-cocO'»-^co O>O>00CO^O>'«-CMO>I^CM T- O ^ CM •«- CO inooo^mcoincM ^^CMO>o>T-T-oo^r^o>^ r«- o> r^ o o o COCOmO^OCR-'-COOCMO ^ CO in •* CO CM iu m OC7>OOCOC3>OCMO Le g Ul sc h <^ •"^ CM C3> O "* CJ> oo •^ 00 CJ >. J -o CO c Q CN ^r»-^cooo^"*oocor>^cocj>co'*incMOcoooin^c2^r*-CM CO o>cocoooooT-coco-<-incM^in^ooooh*incoiocj>h-coinr>-co 00 3 ooocor^c3>cMinT-ooT-cj>CMcocj>T-inooocoin©incDCM-«- CJ ^ppppiv.p"^pCMcoini^r»-inh-;T-;ppph-;pi*-;ppp O •r-^T-^^CMT-^T^T-^'r^CMCMCMT-^'r^T-^^^CM'»-'»-'«-^CMCOCMCMCM T3 CX c: O CN u m C3> o CO o CO CO r-oooocococMi^mcoco^ooooh-oO'T-cor; oocooa>cococo CM COCRh-mCM-^T-T-^OOCOCO 00 r>.cooocj)i^'^r^ o in o oincMCMOT-r>~oincj>o>ooh-cooo CO m o h- CO m "* 00 00 CO CM i^CMinoocoi^cj)cO'*oocMTr«><^'<'0 m "* CO 1^ r^ m COmCMOOOOOOOh'-COCOI^OCOOCOOOO'^'^ T-T-T-CMt-T-CM-'-CMCMCM CM CM CO CM CM CM T3 2 C CN ooT-cocM-^^r^r-co-^cocji^cooo^cor^ooc^inocOf-cMr*- •*OCMT-OO^OOh>-^"*h-COCMCM^O>CO^^r«*^OOCOh-COOO 00 inc3>inin^C3)oococMcococO"<«- qp u a 00>OOCOCOOCOO^r«~CMT-'^h-;h-;^pCMpp^P'^CM"r^i-; CMT-^-T-^-^CMCMCMCMCOCMCOCMCMCMT^CMCMCMCMT-'CMT-^CMCMCMCM schi u CN HH r»-cooo"^cococj>h«-coincocO'^co cooococo^CMr^ocj>cooinooin"*cocMcoinr^i^t^incMCMco CO ino>oocooocoo^in^oc3>inoc3>CMoooooocMCMoor»c3>r<-CM 00 3 co^-r^or^^<»cMh-cocO'r-r^cj>^CMCMT-oocO'*-*incoini^ a> T3 C3)CJ>0OppppP"^CMppppp^CMT-f-pp'^ppp-^ H4 O dddy-y-'r-T^-r^y--r^r-^-r^ddr-^y-y-'r^yr^-r^-r^-r^-r^-r^-r^-r^ •4.^ CX t5 O T-cor^r^inr^cOT-coo>oooo>coo>^ocMcooco '*•«-'* CO T-inin^r^coco^incor*-r^ocoo^T-ininh^i^C3co^^oo coi^co^oh-ooincoinT-oococMcoocor»-r>.T-ooocooocO'»- r^O>^pT-O>COO>CMO>C0ppppppCM^p'^C0T-CMCO^ g J B- T-^^T-^COCM^CM-r^COCMCOCMCMCMCMCMCMCMCMCMCMCMCOCMCMCM o O 0>CMCMCM00'^C0C0T-0>OC0CM^-r^h-'^in^^T-^0>T-^T- r«..oh-'«-o>cOT-coh-oooini^cor^oooocoo^r-o>oo^oococo 00 M ooo^or^-oooomor^ooocO'^-r^T-ocooor^oO'r-oococM 4) 3 co^co^cocoocopppp'^. ppppi^ppppi^'^o^- < ^-r-^-f-^CO-r^^CMT-^CMCMCgT-^CM-r^-r-^T-^CMr-^T-^T-^T-^CMCMCMCMCM 112 o>^^oocoincMr^^-o>ocoT-tO'^coo^ocococooo>oco •"^oco^-cocoocoo^r^^inT-omoor^oincM'r-f^inoo 00 c •-CMin^ocMr^^ino>inoco^-coincococMr^'«-J^i^'^ ooo>CMinino)r*-cMcoo>ococoocoo^h~o>r^coinr-'^cMO ^cocoo>^T-^ooino)0)r^cocMcooh^ooo>oof^^"0>o>in 00 „ mmr^^ococoocooco^ooocooocoo^i^'-oocMcocor*- « >> ^r-r*-o)h-inooo>coo>oocoo>cocoooofloo>o>^o)^cO'«-coco ^3 ci OOT-OOOOOT- O O O O T- O T- CO T- T- CN "^ ^minor^co^ojinr^^incocMcomoocoocoincMh-^CMO CM^r^inco^h-inr^-ocOT-r^inoococM^oocOT-oocoooo 00 S h-^T-cocMinooocooocMT-cOT-incoooococomoo^or*^^ HJ O '«tpppph>pp'r-O^OinCMCMC}>0>00p^pCMT-;ppp CM"^^COcO"t-^^"^CMCM'r^CM'«-'cOCM-r^'t-^^-r^CO'«-^CO^cdcO^ G O OCOCOCD"*COOh-CM-«-CMCMin"^COO'*COCM^CMCO"r--»-C»C3> cj)0"^coocMooinincMcocj>cocoo>r^ r^cMooh^coinooocooom^T-r^cMcorx-in^cocM^coin^oo coT-inin"*co^inincocMincj)T-T-coinocMOcococoooco 00 ocRcor^-cRincMt^^incMCMoo^-^T-^cocooo-t-oincMcoco a> >. h>;h>a>pi^ppppp^ppppppppcopi^ocMoooo J t5 .4-J CO dddddddddddddddddddddd-r^T^dd CO CM^mcMco^T-oooT-T-mcoi^m-^cooor^coococo^oco 00 3 CMC3>COCOin'»-CMCMCOinOCOh^COCOCOCOCOOOO'*CM^I^C3>0 T3 o^^ooin^oocoinoooo>CMooh-ino'*coincooo^^^oo HaJ> O inooo>r^i^OT-incM'«-r*'«-CMincMT-oooino^c3)cor«~i^ >..> CX T-:T^'r^T^'r--r-:t-:T-:f-f-^dT^'r^^T-"i-:T-:T-:T-:'r^i-:^ CM CO UOl •^ CU coinc3>cMi^cMoocoinh*i-cj>OT-coincMCJ>cococj>CMoi^incj> •o 00 CO T-h-T-0)r«-^C0CMh~C0OCMCJ>OO00C3>T-CJ>T-OO00C0T-^ 0) 4> 3 ooo^^-cooocoo>h-h-h-r*-h-OT-|v-o>cM^-cO'^cMinooco 3 C HJ e- CM-«-;^"^ppr-;PppP^"*I^P^pCMpph*CMh--inpCM tg o CMCOCOCOCM^'T-^COT-^-i-^d^T-^CMCM-r-'-r^T-^T-^-r-^T-^CMCO^COCO o CJ ininocj>oocor*-^^cooT-ocgT-oo^coo>cMh«-or*-cMO>^ ^' r^h«-cocococo^oo^o>cococoococMr*-o>ocM^o^oococo 00 CO cM'^incooT-T-o>CMinincMO>o>0'»-f-oini^OT-o>ininin < C3- > h- CJ- > h- C-O -p "^••^^^COCMCMCJ>h*CO"*T-CO<3>'«J-T-OOOh~r« . 2 T-: cvi CM oi CM "r^ CM CJ CM CO CO CM CM ibl e 3 H 113 H 1 u '"^ ibleA • onti nued •i a 1—1 Ul 42 O H-l t5 ^ -c ^ o CM >-i •o s* T3 «0 m I—1 J .2 CN "^ J 2 <^ cC H-1 O 0 CO CM ^ a> 3 0 52 a> "O 0 3 0 —. 2?> a Ul {/J S^ 0 _ 0) 5 0 c IstLe ndL( dLeg c o 0 S Meru CO schiu arpus Oi CM o> CO o o eg p ^ CO CM "* h- CM h- Oi o Csi CO •^ in in eg •«- CM CO ^ "* T— p a> CM Oi 2.4 0.7 2.3 CO ^ CO CM 3.20 in o CO 0 Ol co P CO CM -* CM "* ^— r^ in in CO -^ CM CO 1^ p in o ^ 0.6 CO 2.0o CO o •^ p r^ h- •* d 0 eg ^ CM d p 0 0 CO in co ^- 0 CO in ^ 0 1^ CO ^— h- Csl p 0 o r^ •* in p ^ -^ in '" h-. o Oi m d r^ o o eg p CM 1.6 0.6 1.8CO Ol CO CO in CO CO CO in CO Ol CO Oi m CO -* CM CO p in eg m p r^ •* CO CO o CM eg 1^ 2.0 0.6 2.8 o o o d o m CO o CM eg Oi CO in o o CO o "* h- CO CM CO o o ,^ p in "* CO 0 •^ Oi CO -* 0 o •^ p CM o CO d 0 Tf m CM 1.9 1.2 2.1 o CM O T— 0 0 CO in r*- ^— 0 p CM CO Oi Oi CO m p 1^ o in Csi T— 0 in r^ o CO CO Csi CM m 2.2 0.6 2.7 ^ o> d 0 r^ Oi o •^ CO 1^ o CO h- eg -* •^ y- Ol CM CO m -* CO CO CO p o o r>^ p CJ> C3> CM o CO CO eg •«~ •* Oi Oi 0 t- 2.0 0.6 2.4 d o Oi •^ ^ CO in Csi CO CO Oi 0 0 ^— ^ p CO o ^_ CO •* "* m m -* oi o> Oi -^ "^ 1^ in Oi CO p CO CO eg 1-^ p CO -* Oi p 2.8 0.8 2.0 3.3 CM o r>- in Csi CM m CO •* h- in o o o CO in "* CO 1^ o o oi o ^ ^ o CO o CO CO* p ^ o in T-" p r^ Oi eg p "^ CO 3.2 1.0 2.2 2.8 Oi Csi eg Oi in o o •* CO h-; CO Oi "* Ol cg h- o o CO CO ^ eg -rt Oi CO CO 0 CO -* CO 0 C7> p T— "* CO -^ 2.9 0.8 2.3 3.3 "^ o CM h- 114 Csi CO CM o o in o o Oi "* o o Oi o Oi o o o p CO o o 0 o o '- p ^ o o 2.4 0.5 2.0 2.7 d 1^ CM o o T-^ p o o o T— o o m p o o CO T— o o o o h-- CO o o •* o o •«- CM o o o d r^ in o o p 0 2.2 0.6 1.8 2.2 o o CO ^" o o o in o o p 'O- o o h- o o Oi o o CO •»- o o o r^ o o CO CM o o •r-^ -^ 0 o o p 3.2 1.0 3.4 4.2 Oi o o Csi CO ^ o o o CO o p ^ o o CO o o Oi o o oi "* o o o CM o o csi r*- o o o T-^ CM Oi o o 2.8 0.8 2.7 3.7 CM in o o oi CO CM o o h- o o m o CM o o Csi p o Oi o o Ol o o o "* o o Csi -* o o o ^ CM o o p t^ o 2.6 0.8 2.7 3.4 o CM in 0 o o in o o p 1^ o o 0 o o o o o Csi -^ CO o o CO o o Csi in Oi o o T— p CO o o 2.8 0.9 3.0 3.6 p Oi o o in d o eg i>^ •^ CO Oi Csi o eg V- CM r* o CO CO ^ o o o CM CO eg in in •* Csi in CO •* 0 T-^ m r^ CO 2.3 0.7 2.2 2.8 ^ p o CO o 1^ d m 0 d CO Oi ^— CO CO ^— •* CM o CM eg CO o CO d r^ -* 1^ 0 o ^ 0 •«- ^-l in m Oi d in •* CO o d CO •«*• o 0.6 0.2 0.6 0.9 o •«- CO 1^ "* h- C3> m o d o CM m Oi 0 0 <7> d o d o "* o> Oi CO O C d in T- r^ CM O C d o "* r^ o > CO in d O T- T- o o CO O T- d CO 0 m o r^ d O T- 0 r^ 0 -* ^ m d o "* 0 ^ - " 00 -, a> >. J t5 1.562 7 1.682 1 1.157 1 1.248 6 0.974 7 0.847 0 1.023 7 1.254 8 1.060 7 0.762 4 0.962 5 1.139 0 1.102 1 0.812 7 0.757 8 1.679 0 1.747 6 1.399 0 0.483 1 0.844 1 1.602 4 1.266 2 1.201 0 1.236 3 1.247 2 JS ^ 0.934 2 ^ Q Vi 00 3 155 7 150 0 133 7 380 6 529 0 543 6 901 8 177 5 798 0 256 8 725 8 368 3 545 2 216 7 306 1 757 8 357 8 843 0 (U -o 027 5 812 6 227 0 630 8 HJ O .124 4 .996 2 .910 0 .412 8 CO CO CO ^- eg CM CM eg CM "* -* CO -* CO CO CO CO CO CM CM CM CM Ol CM CM CM CU 00 CM 00 g 135 9 585 1 762 2 395 9 888 3 911 2 915 4 071 8 073 7 838 9 903 7 810 6 398 0 351 6 372 3 597 2 028 0 018 4 068 6 948 7 211 9 474 2 201 7 265 1 275 8 201 6 ^- ^~ ^ o o o r>:S* CoO Tf H3 1.20 8 00 — 177 4 131 3 563 6 914 8 980 4 078 0 066 2 100 5 531 6 879 9 959 5 755 8 064 0 285 6 931 6 747 1 « >» 337 4 920 1 276 0 626 6 470 7 826 2 732 8 450 5 878 9 HJ t5 T— •^ •^ o o o o o o o o o o •o s$ '" Vi 00 3 159 7 377 8 707 9 864 5 532 1 734 8 826 0 865 6 821 3 925 4 755 5 768 6 262 9 620 1 621 1 201 0 835 6 743 9 825 3 746 7 268 5 080 2 O -o 217 8 634 5 427 0 936 2 HJ O Ol Ol CO CO •* Ol CO CM eg CM eg CO CM CM CM CM •H 1* ^~ '" ''" •^ ^ PU •o 00 CO a> CJ 3 1.466 2 1.364 5 1.213 5 1.800 6 1.642 4 1.606 5 1.389 6 1.329 8 1.210 3 1.103 4 1.723 9 1.153 2 1.304 2 1.004 2 1.215 7 1.209 8 1.753 3 0.892 9 2.260 4 1.277 1 1.455 2 1.559 9 0.998 6 0.989 9 1.159 0 i 0.869 9 Td3 BCO- cJn CJ ont i u r-" ?p « 139 3 113 3 132 1 151 8 104 0 792 7 458 0 238 2 085 4 215 4 425 2 027 3 393 4 860 2 236 5 608 6 428 3 201 9 < .574 3 .909 5 .888 0 .722 3 .112 8 .225 3 .571 1 .513 8 H4 2 CO CO CO eg CM CM eg eg CO •* CM CO CM Ol CM CO CO CM CM Ol eg CM Ol CM CM abl e 115 Le g liu m 1.103 6 1.379 4 1.536 3 1.446 9 1.231 9 1.081 6 1.611 9 1.046 6 1.331 3 1.191 9 1.324 9 1.383 6 1.254 8 1.244 3 1.226 4 1.516 9 1.524 6 1.121 2 1.005 5 1.488 9 1.180 3 1.458 2 1.474 8 1.177 7 1.087 2 JS f) 0.963 3 •«-» V) m 00 CJ >^ 990 6 071 5 804 6 950 5 060 9 135 3 818 9 028 4 708 6 914 9 153 2 095 5 787 5 926 5 400 3 •4-» 656 8 732 8 252 9 253 3 092 0 .646 9 .092 1 .740 5 HJ .029 4 .834 3 .987 9 T— JC CuO o ''" o '" '" o o o o o o o o o o o o ••.J Q " CO 00 3 128 3 176 6 CJ TS 136 7 401 8 043 3 432 4 087 4 937 5 567 8 686 7 014 5 957 2 087 9 695 9 430 9 623 5 636 5 293 9 404 5 .521 4 .185 0 .170 6 .445 7 .962 9 .425 1 .584 7 U o eg CM •* CO CO CM CM eg CO CM CM CM CM CM CO CO CO CM CM CM CM CM CO CM CM CO JS CX T••.TJ uo 00 Vi a> 3 1.558 4 1.227 3 1.349 0 1.682 3 1.427 5 1.283 1 1.131 8 1.663 9 1.186 4 1.839 9 1.368 2 1.747 9 1.190 5 1.337 3 1.285 5 1.593 5 1.469 3 1.219 7 1.406 4 1.244 8 1.230 1 1.410 8 1.224 2 1.610 6 1.379 0 U B- 0.969 1 JS uCO 00 Vi 968 0 076 0 400 4 347 6 376 4 733 0 589 4 512 7 518 7 320 8 368 5 569 2 584 0 096 3 551 3 865 6 299 5 469 3 954 1 792 9 993 0 249 1 424 7 4> 436 5 .226 8 .766 0 2 CO CM CO CM Ol CM eg eg CM CM CM CM CM CO CO CO CM eg CM eg eg CM CM CM CM JS ^~ 00 a 3 1.357 6 1.353 5 1.226 7 1.244 0 1.133 6 1.336 3 1.455 8 1.188 6 1.060 1 1.389 1 1.125 9 1.504 0 1.484 6 1.373 3 1.215 5 1.461 9 1.096 3 1.445 4 1.452 3 1.233 1 1.549 9 1.194 4 1.070 0 1.187 4 1.273 0 JS JS 0.849 6 .«.j M "^ HH 00 179 7 150 5 944 4 528 1 362 1 724 7 747 5 808 5 925 5 914 3 202 1 987 2 919 2 746 1 089 8 017 1 787 5 092 8 808 5 775 2 042 8 a> >% 820 2 999 4 308 9 716 8 692 2 HJ t5 o o o TJ CO o o '" o •^ •^ o '" o o o o o ^' ^~ ^ ^~ o o •^ o o •^ Ul O CO 00 3 108 0 538 4 054 6 563 5 024 5 072 5 208 2 007 3 118 7 307 5 909 4 075 0 087 8 427 6 281 0 017 9 033 9 734 2 363 9 634 6 030 8 977 8 336 6 a> TJ 353 4 915 7 213 5 HJ O CM CO CM eg CM eg eg eg CM Ol CM CM eg CO CM CM eg eg Ol eg eg CM '" •^ ^' •^ Ul a. fo ou •o 00 M -1 a> HaJ> 1.342 9 1.396 7 1.372 6 1.132 0 1.525 6 1.084 3 1.191 9 3 1.044 9 1.041 3 1.028 0 1.050 1 1.258 6 1.069 8 1.207 9 1.142 1 1.147 1 0.900 1 1.265 0 1.076 2 1.140 0 1.220 8 1.125 3 1.631 9 1.417 0 B- 1.688 0 1.058 1 G CO 3r d u ont i u ^ 00 M 125 1 120 0 166 1 155 4 122 4 560 8 137 6 116 4 195 6 488 8 355 2 209 9 293 3 620 3 677 4 016 6 506 1 818 2 .835 2 .937 3 .809 4 .371 4 .439 8 .183 5 .377 2 .718 2 < 2 CM CM CM CM CO CM o eg CM CO eg CM CM eg CM eg CM eg eg CM eg eg CM eg CM CM 3r d H abl e s 116 oococor^co<«-CMooT-r^cocOT-coocoeginO'^in^-oooO'^eo •»-CMcoo>coo>inr^coo>^cocgincoi^oo)0)^ooinr^inoocM ooT-cocoo>h-^ini^coo^-o)o>inoooo^-cocM^oooocM CM CO -^ "* eg T- eg ^ iu m Le g OJCOCOOOCOCOincOT-CMOOCMCOCM JS oi •<-» CO m HH h^-Oph-ocMOT-coococRininr^cMT-cor-CRCMT-^rcNo- o> o o>oor^egoocoh*oococMcocMegc3)cocoegT-r^incor^coin'* 00 oiocoincoinoocoor^^cocoO'T-cocgh-cO'r-coinco^cg >. a> •4-J ooooocMt^T-^^-opcMT-ooco^-mppinininincoininin HJ d-t-^dr-^d T-OOOOOOOOOO JS CoO -4.J o CM CO O h- ^ h- CO^ooi^egcj>ooegcooocM^i^^cocM"*-i^^- ^ CO C7> O r^ 00 O C7> CM ^i^mcocomh^cMOcoococoh-coi^T-^ro. 00 3 o> o a> 00 ^ 00 r^ ^o)inoocoT-c3)0>r^^in^inoocO'»-ocoh- a> •o CO CO in -r- CO O CM pT-h^0OpT-0O0Or^T- h-r^oooT-cocor>^incMor^oco<»incj>incoegcMCMooincMco 00 CO ^'«-O00'«t00CMCMC0C0T-00inOCMOOC0C0"^C3>-r-C0'«tC0ini •«- m m ^ n (U cocoT-cg^T-coinr^cOT-oin oocoh-t-CMinT-cot»-egincococococo<3>c»"*cor>-o>oeg"^o coh-ocg'*h-oocMCMinc»inT-cOT-ooooro-OT-in''^T-ocoin coocoh-ooin^coocMor-CRcoT-cRco^mcMh-oocooo^o 0"*"*cOT-incgincocococMC3)^co^T-co^oo-t--«-r^cj)egeg •r^^T-^T^^T-^CsiT-^T-^T^T-^T^d-r^-r^Oi-^-^^T-^d-^-^-r^T-'dT-^T^ hLe g chiu m •«-» Vi "^ HH COOOOh-^COCOCOinoOCO'^T-T-T-T-COCMCOOh-r^COOCMOO 00 •»-r«-r«-inc3>cocMinoo"*inT-o>cgr^i>-T-cMO"*eg^cMcoin-«- a> >% ocoi^inor^T-inc3>cMCMcoc3>c3>coincj>^egcMin^egc3>h>-eg HJ ppi^CMpcsji^ppcsipppT-CMp^ppp^inco^inin TJ CO ddd^df-'-r^dd-r-^'r-^ddT-'T--T-^dddddddddd u Q oco^c3>"*cocoooc3>cooegr*-CMoo"*ooT-^CMr^coc3>oooco (/> h*'«-cococMOO>i**o>o>oooocoO'«-inh*inincoincoinf^cor^ 00 3 r<-oinooh-OT-coo>cooocococMco^T-^c3>r*-oooooocj)coin 4> T3 oinoino^c3>i*«-incocoegoo^c3>ocj>oegh-cocj>inr^o>t^ H-1 O csicMesicsiesiesicocsicsicsiesicM'r-"csicM"^'f-^CMCsiT-^-r^T-^cMT-"T-^^ fo inco^-t-^^cDmor^mcoh^r^h^r^coT-inoinooooo^in oocDcoi^"*o>oin^o>o^T-cMcocgT-coinooco^incor^oo t3 00 CO r^coooT-cor-^oincoincMCMcocj>ini^o>^o>'»-ocoo)T-oo a> a> 3 ococoeg'«-^^oo^T-cM^C3>"*h-ocM'*h-cocMh*-o>T-ineg 3 H^ & T-:T-"T-'T-*T-:-r^CsiT^T^T^T^T-^dT^T^CsiT-:T-:T-^T^T-^T-^^.r^.r^.r^ G CO -Uol *€ o O CJ r«.inooin'r-cMa>ooinoococgcMcoa>ooa>o>T-cocooincO'T-a> oocooo^T-'^ocO'«-cgo^oooooinocoT-o>coh~CMO>coo cocoooinoor^coco^inh-OT-cOT-oooegcMOO)inT-incor*- < CMcoor^cocoppi^ppppppppppppT-oomooo CNiesicsicMesiesi^cocMCMesicM 117 able A. 1. Continued 3rd Leg 3rd Leg 3rd Leg 3rd Leg 4th Leg 4th Leg 4th Leg 4th Leg 4th Leg 5th Leg Merus Carpus Propodus Dactyl Ischium Merus Carpus Propodus Dactyl Ischium CM •^ CM •^ eg 6802 6006 9656 0235 1.7342 CM 7178 1.5706 6733 1.1369 1.6906 Ol ^~ CM o CO 5850 4708 6480 9829 1.7875 CM 8266 1.8685 6031 1.2618 1.6098 CM •^ eg •^ 7121 4830 5197 0705 1.7470 eg 8293 1.5960 CO 1514 1.0798 2.0874 CM ^~ CM ^ 4262 3968 7959 1932 1.8578 CM 6471 1.6304 CO 0824 1.3015 2.0143 CM ^~ CM CO '" 8293 4499 6460 2424 1.7984 CO 0318 1.6999 2204 1.0723 1.6543 CM ''" eg •*" 7966 5754 6063 0225 1.7818 CM 8177 1.5257 CO 0886 1.1921 1.7633 CO •^- "* CO CO 6376 .8717 5557 ^~ 4519 2.4109 3891 2.0262 1105 1.6102 2.2536 CM CO '*" eg CO 2379 .5174 3358 o 9145 1.9825 0885 1.7808 1548 1.1418 1.6837 CO eg '" CM Ol 7125 .4802 8990 ^~ 1862 1.7012 7254 1.7492 1485 1.1720 1.9809 CO CM CO 9476 eg .0542 CO 3422 ^ 0389 2.0104 0692 1.9046 6337 1.2256 2.2084 118 CM CM o '" 7835 .9772 7968 o 9150 1.0477 0388 1.0706 1397 0.7263 1.1150 eg eg ^~ o .8236 .1426 eg 0310 9541 1.0879 0763 1.3460 4745 0.8002 1.0293 CM eg ^" CM .0781 ^ .0118 7571 0192 1.0753 1676 1.1027 3308 0.9906 1.1088 eg eg .9605 ^ .0074 7434 '" 0854 1.0380 1223 1.1568 2653 0.8180 1.0640 Ol •^ o ^~ eg .0694 .0749 5206 9434 0.9835 4495 0.9940 6728 0.9666 1.0749 CM CM o o T— .7203 .8926 5971 8992 0.9833 1571 1.2323 3849 0.9421 1.0904 '^ •^ o .7470 o .7790 3814 8332 0.9668 5520 0.8991 8500 0.8144 0.9857 CM CM o .8727 ^ .0171 5960 8975 1.1966 1058 1.0582 0739 0.9234 1.1546 0 0 CM 1.99 ^" o .8556 o .8187 5418 7516 1.0698 1.0319 9678 0.8055 1.0459 CM o CM ^ Ol .1748 .3067 7907 8881 1.0940 4793 1.3728 2189 1.1522 1.2081 CO CO '" eg ''" eg 6542 .5673 6356 1391 1.2790 2428 1.5760 0362 1.4302 1.5921 CO ^" CO CM ''" CM 7357 .5921 3819 1976 1.6696 2944 1.6983 .2320 1.2732 1.5659 CO '" CO CM '*" eg 9700 4636 2835 2539 1.4689 3693 1.8762 .1923 1.2692 1.5948 eg CO o eg ^~ eg 6003 1297 6806 6741 1.5500 9283 1.5197 .1459 1.2290 1.3141 Ol CM '^ eg V- CM 5721 3776 5196 1623 1.4577 4402 1.2085 .1614 1.0868 1.4898 Ol CM '" eg •^ CM 2129 0660 0141 0508 1.2002 4335 1.3730 .7449 0.9461 1.3415 OOOC3>CDCOCD'*T-CO^CO-«-p53-20 oo^oo'^mo'^'^ CMr^co^inc3>cocoincj>r«-3-in02hO a COCDCOCDCOCOC0CJ>00in ocDincocoo)coooi-incoo>pop5 3 coo^CMh-^T-o>o>r*^in COCM^OICOCO^^CM'^ Le g •»-CMCMCocoT-r«*or^inooco^coo) JC JS •r^T^'r^CMT^T^T-T-T^^-T^.r^T-^^d CO ^r^coi^cor^incooor-i^cgco5rpcM ^or^incor^or^coo ^oooocDcor>-T-inininini^cor^o> 00 ocDCMi^cor^o>oinin dooh^ooT-cocoinPooino>h^oa> a> >» ^T-OCO^T-^COT-T- HJ r^inooinr^ococ»i^'«- ^OOCMCOC3)C3)CO^OC3)^PPPP^ ^ CO o o o T- o O O ^T^d-r^-r^-r^^'^'r- •4.J Q Or^COT-C7>^CM^COCO ,_.,,_incMCDom^mco. ^ ^ <3> O 00 O CD CD CO ___ _ .mi^CMCMCOCO 00 3 COCOh-C3>CMT-'^COC3>COCMCOh*-«-CO O 00 T- o CJ i^O"^coT-ooor^inco NdJO^cor-cococ^or^cgincMc- o •o ^.r %-<• «_« ••# I'" l"~ \Ji \-^ MJ •^- SJI \y UJ 1^ V^ 1—^ t'/ «L> ^— ^ W.* »*' *" *" "^ ""^ H4 O ocMoint^i^ocgco^ 0>CDinCMCMOC0CD'«-'^00000)0>00in JS CX cg-t-^esi^cMCMcgcMcocococMCMcocMcococMcocgcsicsicocsicMCsi ••-» Tf oUl CU oc3)^cO'*coT-^r^oor^h-incMCJ)cocO'^cocM"»-incDcO"^ ^incoincocor-oJcgr^coegT-inooinr^coinincoocM'^T-oo 00 CO o>coo>r^r^coocM^coco^o>cO'^^^h-cD^h^^h*^«no> a> 3 ppp^coT-co^c3>i^cj)^'«-coincoco-«-oop^pppp^ U T-^ -i-^ d CM T-^ T-' T-^ ^ ^ -r-" -r-' -r-' -r^ T-^ ^ •«-• T-^ ^ -r-' -r^ ^ T-^ ^ ^ T-' T-" M cBO- ^ o i^"^inocM"*^in"*'^h~ooco^oocooocoooocMOincMincM C»-«-CJ>C»O00CMCMCDf^00T-T-'*OO"*C3>C000C0CM0>0>r^O cDr^incDO>'»-co^r*-inr^cDO)T-coino>cO'«-r*-coo>^coinr^ pegpp"^csicMp"*pinT-;cooocMcoT-pop"^pp^pp |s T-^cg'^esicMegcocMcocococMCsicsiegegcocsicocMCMcococococo' r«-"^oin^o>cor»-ooi^c3)OO^CM'«-c»cgi^'«-cj>o>^r^oocM o>^coocDCMC3>o^h-ooocj>oini^inoin-r-inoocMooo"* 00 G ^ooinoor^T-inoooh>-oo"*coco^cocDo>T-in^'«-oo^r»-cj> a> E oo^cj>cj>0"*T-o^coT-egcM^inco^coocMCDcocococM CM d^T^ JS o CM ininocj)^coc3>cgcM"*r^o>inoocgcoooocooocococDinooT- c}>egoocoeginego>r»oooooocMC3>r^cor>-cot^r^coooinoo^o 00 —, cocoin'^incoo>'«-cococOT-oo>oococooi^coo>o>'*-co^co r^c3>r»-cooooOf-r>*-egococj>C3>ooooocMOc»r^i^o>cDT-^o J t5 dddT-^ddT-^d'T^T^T^ddddr-^^-r^dddd^i-^-r-T-: cn '^ "^r*-'«-cD"*ooh~ini>^T-"*coh-^oor«-r*-in^"«-egcj>incj>CM CO co 119 CMOOOOCOCOCM-r-COi- r^oocDCMoocMino>ocoo>oo>oocDT-r^ 00 G inoco^^comcM- cgr^^-coTT'T^oinincocgT-cocM^bicMr^cgiO ^CMoocococMincocgo-«-cD^oc3>xhin n ppppph»o>^— O)O>COOOCD^OT------—-- ^r^oo^-o>e. _ _. ^ ^. .^g -„ .^^ ^ ^ •«-'»-cMT-T-dd'T-ddddd^^'«-ddd^d^-csicNicg^00 o r-^ inT-h».cD'«-T-o>cgcocMT-oococoo>cDcgr^^ino>OT-o>cD ooocoo'^cM'T-'«-a>oo^ooincgcMt^a>r^^oococoh>-i^oocg 00 — ooh-h-oooocOT-oO'f-ooinT-coooinoooocgoininf-T— CO CMppp'^a>^-r^a)'r-cooot^T-cj>cDr^cDooo>ooa>inppesi ^•»-T-T-T-d"<-'T-^d-r-"ddd^dddddddd'«-^csi'«--<- oco^r^^coho-ooh^ooegr^inT-aior^ooo'^^co^ocgoo CO eginh-oin^ooo>OT-coinr^^coh*ocD'^'«-o>r^h*-cocooo 00 3 CMo>^o>co^egcD^cor»-T-ini>^o>T-r^^cocO'»-^»^ooi^o pppp'ttppcgoocgcoinr^^i^r^oooo^-pppoip^ HJ O CO •^ "'t ^ CO T-' CM CO CM T-' T-: .r-* ^ Csi CM T-* •«-' •i-^ T-^ Csi -r^ CM ^ in CO "^ CU cgco^oiOT-ocMco oincoooocj)CJ>o^cocooocMinT-cOT- incooinc3>eg"*eg^ 00 CO oioor^^^cRcgcDCM CM<»CJ>OOOOOCMr^OCO^-<-COOOO^T- 4> 3 CM oo C3> C3> in r^ ci ooo"*^cj>cj>cj>incM"*inininco^cMoo " CO ^ ocor*-oo"^coooc3>oo^o^T-inooin .C CO T-' T-: d -r^ ooot-T-ooT-T--»-t-f-egcM^CM ^ CJ cO'*h-ooooooor»-"*oococgooco^cMCDcococ3>egcococ3>oo> I^T-inCMCMCOCMI^inincDCMOCOT-COC3>a>l^-^OOCOCDCOCMCO inCDCMT-^OOOCOCJ>CMCOT-COOOCOCM^OC3>r^h-OOOCMCOCD P'r-pppr«.pppcMT-;cocor*-coooT-ooo>copppppp esi'^^'^COx-^-r^COT-^CM-r^T-^'r^CMCM'r-^CM-r--r^CgT-^esi^in^CO •^T-incgcoinococoocMCMOcgcOT-ocgco^T-oocor^ooco "*in^oocMOOh-oocj>^|x-T-c»cooo"*^ooT-i»>-T-cgr^r^ 00 G co-T-oocRT-ix-r-cor^ooincRCM-^-ooincgcM^mcocDOoooco CMppppr>>pCMppph<-h--ppppppcOC7>CMCMCOOC7> JS o •r-^'r^^'r^T-^d-'-'-r-^ddddd-'-^T-^dddd-r^dT-^oioicsiT-^ ^ I-H ^incoooT-'^C3>inooocoegcj)co"*T-T-cgcD^T-^oo^T-co egT-cDegcoT-cg^oin"*cDoocDcot»-oocj>ooincj>cDh-cgT-^ 00 » egincomcor^ooor^cooO'T-r^incoco^ocDCMrx^cooococD coina>in'^pppppr^ph^pppppcoooh~a>inocoo o o'T-^dddddddddddddd-r^oi-r^T^ r>^cj>"^coego'^ooocgh-cD^^co^O'*^ooc»CMcoT-oin CO oo>T-co^-T-co^coo>oo>coino>^T-o>i^'d-cocDeginT-o> 00 3 r*-c3)CDooco^ O CJ> CO !>»- cooo^^-T-o^coo>T-incoT-'*-ocD^T-coo>r^in ^ ^ 00 o ^coinh>ini^egoooinooc3>cj>cooocoi^coincooo^ t3 00 CO _ _. ..•^oocDcocRincgoocMcoh-co'^in'^CM'^^T-ooin"^ 0) 3 oo cDCDincor*-o>coo>o>pf^r^pT-;ppooo>CMO>coo>'*-oooo in CM > •i-'T-'T-'^ddT^dddddt-^T-^dddd^dT-^T^CM^f-^ »- O O m i^" 120 ^ O O CO CM ^ r*- CO 2°*I:;I^. ^*^'''«>"«- r«>i^r>>h-c3>eginegcMcooooooo ^ oo •- CD 00 co^T-T-T-cMcor^^cjioooooo ^ CM CO m '>^ 2S?iQ525^ cDocooooo>oocoa>inoooooo •^ ^ CM CJ> CO ^h-inoi^coT-ommoooooo 00 3 inincM^cocooocoegr^egr«-^T-h-r^ ^ CO C3> CD 0) "Ta HJ o ^pT-;pT|-CDCOinOCO<»"*CD'«-C3>T- h- h- in T- J3 CX COCMoicMCMes|CMCO"^CO-r^CM"^^CO"^ 00 1^ p T-, '"I- uo Csi d d d cocomegco'^ooegr^ooooooo T- •«- 00 00 •«-coc3>incoinininr*-ooooooo "Tt "^ o o 00 CO c3>'»-ooc5>in^coooinco-«-o^"^co ^ O CO CD o 3 inoococoin^r^o>cDoor^cocj)T-T- m "* o o H-) CM CM eg CM r-^ d d d JS CBO- -4.^ TT u T-i«>.inin"^coegcoc»r>-oooooo Oi Oi Oi oo coocMT-h-ococR'^egoooooo 00 M COOC3>inT-COCO^COOOinCDOOC3)CO-<- 00 C3> "* O CMCMT-;ppph-.ppinC0O00h-;l^r- •«- T- in T- 2 cocMCsiegoicsiesico'"^"cococsi"^cococo JS 00 r^ o o CM d d d oo^T-inO'»-h-ocoooooooo h- 1^ O CO a>c3>in'^cocoh<-c3>inc3>oooooo CM CM 00 in a or>-h«-cgcMineg^coi*«-^CMO>coin-«- CO CD CM m 3 inoO"*T-coT-inr^^ooT-cMinr^in CO CO o o Le g -r^ d d d JS •4.J 2Mc> '^ I-H oo^T-"^inr«-^0)co"^ o o o o o "* CO CJ> CO comoocMcoomor^o o o o o o o CO T- CO 00 lnco•r-o>oo^ooT^ooo in r»- C3> CO CD CM 00 CM ^ a> "^^ OOOh>OOOOC3>OCOCOCM CD p p O ^ O CM O O HJ f-^dddddi-^^T-^^ T^ d d d 'O oCO ^cOf-r^oooooT-cMco o o o o o m T- CO CM CO o o o o o 00 3 Oh^COOOCOr-incOCOO) o eg CM eg o m o ^ h- a> •o oococomr-or^-oh-h- CD in h- h- h- •* T- in o ^ O inc3>coc7>ocMCMa>ini^ Csi CO Csi Csi Csi CX csiT-^T-^^CMoiesicsico'cM CO 1^ o T- Ul m oIM CM d d d T-cocommi^co^oco o o o o o 00 CO CD CO ^ocMCJ>h-incocMooin o o o o o h- in m o •o 00 M cDCJ>cj>CDincoocMin^ ^- CO CO ^ CO ^ CO CM in V 3 incj>cj>egT-cMCMini^co eg C3> in 1^ CO CO CO o o a3 > hJ G B- T- O O T- T- ^ T-T-^ CMT-T-T-T- ^ d d d CO 3r d ont i CJ CM'O'Or-CDO'^-CRCMin o o o o o in CO CO oo r-i o o o o o 00 M oocooocDor*-co^o>co 00 o O "* T- CO ^ ^ ^ 4> COCOC3>000"*incOT-C3> < 2 0>ppppp^^h^^ P "^ p CM CM CD 1^ m o o csiT-^^esicsiesicsicococo Csi ^ cd CO CO p CO o •<- Csi d d d 3r d abl e s H 121 cor^cMt^o>i^^ocg^'^'«-ooo>cgoo'^0)0>coeg^inr>-o>co •-I^^^CMC0C0^inOh>-0>"^Or«-C0C0C3>OCMC0"^0000CMO c 5; ^incOCOCO'^OOI^OOint-COI^O)CMOO)00"*COCMCM^^T-00 O "^ ppph->pph..o>ooT-'T-h~T-a>ooooa>a>oor^pppppi^ .52 X '^•^-•^o-r-ddddy-'-r^dT^ddddddddddddd OT-5CMh-T-coT-Tj-ino"^'^coinooc3>i^cor^oc3>coococM G .c CM^^T-C0C0^C0C0000>0>00'^'«-00CMC0'«-C0'^CgOO00^ 0 r*i^cDcOT-o>r^o>oo)ooococof^ininoocDO)oocgcooooco Sb ^CMh-.pcM^pcMcoooo^inoocor^inT-oopcMpppph^ a> ac> '^^"^esicdcdcdcdcd'^^cd^cdcdcd'^^csiesicMcdcMCMCMCsi H J OOI^^r*.T-h*COCOCDOOOinincOCOCDCMOOCOCDt^^OO>OOCM 00 T? h»o>cO'^^inininino>cDcoinincooT-o>incDT-'^ooT-cMO> (U CMCM^oor^cM^cDCMT-egegi^coco^r*-cocDOoocMcOT-^co C/a3> (U p^, ^ppcsipi^r^ocMOCM^oooor^pppppppp •r-T^T^dd-r^T^dd-r^T^d-r^y^y^d-r^dddddd-r^dd m incDcgr^ooooooomoooor^r^r^ocof^T-oocoinooooooo) ego)h-cocoo>coi^coo>oO'r-incMcOT-o>cDcoino>'^cooo^r^ coinoO"«-cooooeginoocoinoo"*ocMr^cg'«-T-coc3>oo^^ CO PH T--p ooc3>r^c3>cococooc3>'^egcoa>-^cooincMegoo^C3>c7>incoc3> O •«-c»^T-coc3>CM^cooooinooococoh-coi^ooT--<-inoocOT- CO cocj>coT-cgo-«-CM^^r»-^CMO)c»inoo^or^cDT-oooT-co 'B - OOr^inT-T-cOT-CMCMOI^CMC3>CM'*"*inCD^COOO'*-^in'«r- *< CX CM •o o r-> CU CDh>-'0"T-T-OOOI^OO^OOC3>OOO^inh-CJ>CDCOC3>CO'*OOCOOO ^coinina>a>inh«(3>coh<-cocoooooc}>coh>ocMoor«-ooroco^ ooi-cOT-oooo'*T-coininr*-ooT-0'«i-h-cooor*-inc»T-oocoo cDoocMO^CMOT-T-T-oo'^cgoot>-co<»oocococOT--.-CD^eg CM CM CM CM "*egincooooocj)r^cgegcginooo^coh-oh-egor^h«-cor- egi^-«-"*cDcoinoc3>egor^cMor*-ooc3>cDor«-egcj>cj)incoin 00 —. ^inoococ3>r<^cooo'^oo-<-r^ooooincDCMC}>^incoinoocoor*- 4) >» pppp-r-;ppcMT-^pppppcMT-^^pppppooooT-iv. T-^y^T^dr^dd-r^T^-r^r-^-r^'r^T^T^T^T^-r^r^T^ddddT-^d 10 •"• coinincoococoa>coc3>cococooo>cMinoinocoT-or^oocD c» coocD'^cj>inoi^^-«-cococMC3>cocoT-oooincoc»i^h-T-^ 00 3 CMr^co 122 coocDOt-r- o>o>oo>i^cDin^"«tcgcDin^cO'^OT-cMCMO_T- _o •«- o "^ -^ cocDinr^inr^inc7>ina>cMcococ'"""•-' ._- ____-..__--o _o C—D o^ coooo>oocO'«-T-egT-^inT-cMcDcoooo^inT-^-coinco^in c3>incMT-cgoioor^-«-egincococo^eginr- f*^ r^cMcoooT-cMCDinc3>ooooincMCDcoh*r^incoh-oocor^i^co cooocoooooocO'«-ooooooooi^incoi^ooinoinincMor^oo cococ3>cocgooincMCMOco^co^"^in"*i^h^egcooinooegr^ pp^ooi^r>-t^h>-c3)oocD^oor^ooooo)pr^r^h-;PP»^pp c/D Pu dd-r^ddddddddddddddddddddddd CM cocj>oooor-r^inoo"*^coo>incj>cooocM^^i^cOT-"^in incoco^r^'*cgin"*cDC3>c3>egooT-egcj>inegc3>coooino>'«-co r^incMinoocM'^ooooc7>ooooooooinc3>CMOO^^_ ~CJ > 1^ —o C3-> -r*. inT-cocMT-incMO"^T-o>oococoinin CM CO T- CM eg CM T- oo CM o o d T^ OCMCOOO-t-CDr^h-OOCOCgcO-^OOOOOOCOCMCRCRCMOOOegOt^ cocDinr>*'*0)h»-in<^c3>coinoo^r^c»^'«-cooocgh»i^C3)ooo oocD'^cgT-a>incooincDoocDi^cocDC9>oocD'<-^oocoocoo i^co-f-co^"^inegcDcOT-c3>^r>-r^cD"*cO"*coco^O)CMcgoo eg "^inooT-ocD-'-inr^cOT-oooco-»-oococ3>-*coco^egoo)Tf o>r^cDcO"»-ococM^oocoo^^i^ooinincj>cDoo'^c3)i^in 00 -^ cDO"*inooinin^CMegoT-ooc3>"*o>ooh»cDi*^T-incoin^co 0>pCMppT-^PpT-;pppppT-CMppph>-ir^Ppi^h-T- JS ^ ddT-^dT^^"ddt-^t-*ddd-r^^*"«-'T^i-^T^d^ddddT-^ i/-> ^"^ inoor^inr^^r^OT-OT-ooT-coooooor^-cMcoT-r^-ocoo CO ocDegcocoooT-inoo>T-r>-cocoooT-egt-r>-oooino>egooo> 00 3 oo^oO"*incj)T-^i^egor*-r*-egcj)ininoo"^^cj>^cg'«-co poin^pr-;egT-^pT-p^f-cMppppppppooo)oin J O cdcd^cdcdcdcdcdcdcdcMCMcdcdcdcd-^csicdcMCsicdcdcsicdcd _C CX CU CD 00 eg CO -^ ocg"*cooooo"^ooT-ocoooocMCDocM^oegcD CM o m O C7> ^incocj>"*h~oocoo)coeginoooT-or^ininooc3> 00 CO C3> CD O CO ^ ooo>egocMoc3>T-ooincMor^r«-cj>oin^cococo O 3 oo CD '<- O CO •*CDCOr*-"*^CMCMCMr^ppCM"*CMCOCOh>-incO0O g CM CM T-^f-^T^T-^T^T-'.-^T^T-^-r-^T-^oi'r^T-^'r^-r^i-:,-:^,-:,-: JS CO o ij^ o ininoor»-coo>ooo>coincMO>o>cgcocMO)CDT-cMr»-ininT-cM»n cDOC3>ego>oo^co^cDCDco^h-coincgh-coooh>-oocDCDr^cj> U PH T-ocRco^^ooego-^or-cDooi^cooocgT-T-cDoor^incoco H^ 2 CM"^CDCJ)'«|-Oppppp-«-;pPp-«-;p"*Pp'^"^"*in"^p cdesicdcMcdcdesicsicdcMCMesicsicMCsicdcdcMCsicsicsicgcdcMCMcd 123 O -^ -^ CM •«- 1^cocMooo-^oiincoco^-mor-coinr^oiincM ^ CO -^ "^ CM ^ "^ c3>CMcgT-r^c3>oo^oco"^^egT-<50)or;*-cpco OO CO O CO CM "* r* CO 0• 0 m S- in r^ CMOOomr^minT-oocor^co f^ m C7> 00 C3> 00 a> r^ ocM'*cpr^r^pppi-^PP oococoo^coo-«-cgcoh*r^ocDCJ)i^cocM^O)egcoo"*CMoo «5tQT-r^'^0"^cor«-inc3><3>"^T-incj)^inocDinoT-ino G JS cocoo>^r«-^co^^eg^r^oegcocoinco^o>o>cg^r*-^co O ^ ppp^p^O>h-ininCDCOC0CMCDT-O)^'r^pr**;^PP'^. P Vi csicdcMcdcdcdcdcsicMCMCMesicsicdcd"^'r-^csiesicM-t-^csicd-«-esicM CJ uc H J h-init-«-oor*-incj>cj>CMCDin^eginT-coooocD"«-CMegcor^ cocooegcocMincoocROCMC^^CMcocMcgoor-cDr^ppj-co 00 T3 ^oo-^cRinr^cooco^omcoT-ooocRoego^r^hrpr- (U a> ph-r<-ppr^ppppr--T--^pppr«.pppppr^'*pp (/3 CU dddddd-f-r-^ddd-T^ddr-dddddddoooo f<^ "^h-cR-^oooc^T-r^^m^co-r-cMoocoegT-cooopT-cocDoo cOT-coO'^>«-cO"*oooc3>c3>oor^oo>"*cococDr«-cgoocoo>^ 00 "O ooinoor^cj>ooT-^co^co^ooo^"*r«-i^co-«-coi^coooo|in V a> t^or^cj>t^ooegoocooocDh-ino)OOco^^^co^pp^'^ C/D P^ dy-^dddd-r^dddddd d'r^T- ddoooooooo 2r d An t *^^ cMOoocDoegh-cDC3>^co^h-oocMCM'«*^'«-ini^oooego^ oocj>cocDO>cj>coT-cMCMcoh-oin"*coooocOT-incO'^T-inc3> CO I^OOOOCOI^OOCJ>CDin"*CMCOCD^CJ>OCOCMOCOO"*COh-r^ cocM'«-cO"^'^^"^T-"^"*cOT-inr^cj>oocgT-c3>T-cooooo Ol S p/ D T^T-^-r^-r^T^f-^t-^T-^T^T^T^-r^dT-^-r-dT-^T-: T3 u uo CO CU • •-* coin'«-cDT-ocMT-'^cOT-r^T-cocMooooincMT-r>~ego>ooo>oo T-ooT-o^inooooocDOcocOT-r^oooh^T-ooh-co^o M ooooocDC»h-egcoinooinin"^cgcor»-cMCJ>CMincMr^r»-^coc3) V.ed 3 cocomin^cD-T-inT-cgcocoo^cD^T-ocMppppppp :^ T-^-r-^-r-^T-:^^"CM'r^T-^T-^-«-^^"r^"«-^'r-^esi-r-^T-^>r^t-^T-^d-r^d-r-^d ar p u f^ u egcocoeginoin"*c3)ooco"*^oocoT-cj>c3>CMegincooooT- T-oooocM"»-^'*-cj>ooo>cO"*cocor^-«-'«-»ncMooo^OT-or»- CO egin^eg^cocoooinoinocj>coinT-c3>^r««.0'»-oocMooeg 00 3 cooegcocgincococoino^ooocgcoino^h-ooco^^cDO) O •o cooocoppegcM'«(t"egpi-;ppppcsjpph-.ppcMCsih>.poo HJ O cdcdcdcdcd-'O'incdcdcdcdesicMcd'^'^cdcdcdcdcMcdTrcsicsiesi JS CX m uo Pu cj>cocoT-ooocMC3>cj>ocg^-ooooocoinf^inr^T-egoooococo cDincooocoinr«-coinT-^egoooocor^cDino>^egT-o>ino>T- T3 00 CO cocooooooocD"*"*ooegoocj>r^^h-cO'»-r>*oooooo^cocg U o 3 ooocoini^ooooo)h^ininin^-cj>i*^cMpT-^ppooT-h«»ooo>h^ g HJ B- T-^CMT-^^^-i-^CsiT-^-r^T-^v-'T-^-r^i-^CsiCM-r^CsicM-r^'r^CsicMT-^T-^-r^ JS CO is -u» o o U ^-o>cocO'«-ocDr*-»n»ni»«-cDcg'r-o^cMcoococoo>^o)coeg ^ coinoococj>cj>cDegincooocoococoinh-or»-ooT-cocor^cj>eg 00 CO ^r>-ocj>cDegh«-CMego>T-inoooo^CMcoegh-egegincoooo^ cot-coocDco^ego>o>ppp'^. p^-:^^-;es^ppppT-cD^ < CJ2 esicdcsicdesicdincdoicsicdesiesicdcdcdegesicdesicMCsicdesiegcM abl e H 124 ^ti2r":"''>'^<*5 ^fCM^CDcomin^cooooin^mor^ in'«-cM^ooo)'^coinr^ iGS5S2'^f^'^«*>^''^oi"^CMininr^ oo^t^r^o^ocooo c JS Si^5^S50 cocj>i^cj>^^^ocj>coooc3>o>^coin^cocMegcOT-r^r--T-oo ^h«-o^inooininin^egcMO>r»-ooocgcDCMOoo^cj>©cocg 00 •o egcoo>ooint-coin^coo^r^^^o>T-coi^cDCMcoooooT- U 4> p^lncDlnlncocDlncDCDoo^>-^<-r^l^cD^^^ooo>^pppp cn CU ddddddddddddddddddddd-r-T-ddd r*^ "*cor^coocj>r^oooocDoocM"^co^"^incMi^ocDOT-cMin^ ininocMT-cOT-inoooocM'*T-incMOO>oooocoint-o>o>o>T- V^ CJ15 r^^co^^r^cgocoocoinocor^cO'T-in^in'^ooocoooo ^pppppi^'«t^incor^t^h-ooh-coppppp'«-. ppi^ in PH ddddddddddddddddddddd'T^-r^ddd CM "*^egi^cMC»coinr^oooincoc3>cO"^co^cDO m CM m eg o> C7>inoincMC3>ocj>oocococMC»coocMh-egoc3 > X CO cocoi^egcDT-ooo"*co^O'<^cj>c3)ocor*.egcMinoi^inr*-cCO "^ 00 CD CM t^ o coinegcocMT-i«^ ir> •"• •«-eg"*cocj>r^cMh*coooooincj>o CMCDinT-CMO)CO'«-^COC»l^ co^^inc3>T-T-i«-cor^^ocoo CMOT-'*-^oo>'*-incoooT- 00 3 •«-inooin^cocD^egooo)coh>-^ ^eg-f-CMT-oooco^cMOO ococgoocococDCJ>'*"«-^cj>h-r«- i^coincDr^^oincMoocj>T- J o cocococococO"*cocO'*CMCMegcgT-cMCMegcMcO"*'^' 125 oocMr*-oocOT-^coo>^T-coh-^cgoocO'^^^'^'»-<*'P ooinincDoocgcMT-r*.ocMcocg'3rcMOoococoococoincM^cM ^(Oi^OT-'^oc7>ino^c7>-«incD^'^oo<^^c3>inT-r^incsi o ^ ppppppopoO'v-inhoh-cDoooooopi^ppp'^p J2 X oooddd'T-dT-T-^'r^dddddddddd'^'-^-'-''- moocMocooocMinT-co^cMinh-cooor^'^OJcoincot^ooo r"2^'^5^®"*'''Oeoc3>cocoo>r^incM"^cococOr-]0;CMOoo r>-0)egcocooO'«-cocD^coT-cor^"^^oooocO"*cocgcocj>cM T-;pesiph*.pi^o>T-oinoo^^oO'r-eg^pi*-;pppp'"*:P CMCMcg-^csicsicdcM^'^'^csicdcdegcd'^cdcdcMCMcdcdcdcdcd ^cocoinoO'*-oooooT-o>r^r«-T-egor»CMoo^oor^ino>ego ^oegcgincoegcooo^i^o>egcor«-T-r«-oooincococoooflO cDoo^ooT-oo'^'^or^ocoooT-T-ooi^^cgcocoi-cMcoo C/D (U ppppppCMpCslOCMOOO)OOOOppCsiph-;p^p^p ddd-r-^ddT^dT-^T-^T-^dd'«-^d'»-'»-d-«-dd'r^'»-^'«-'«- m T-cDcooor^egincMinooinin"^CM'^"*"^oo<3>oincoco^ooin in'r-"i*ococ3>in^cDinococMcocj>inc3>^-<-cooocMinT-ino 00 "O O'r-OC0I^CMC00000in"r-"^CDh~OCgC0CDCJ>CJ>0>^I^T-|^00 ppP'^h-^pT-h.OOCMOOOOOJh-OOT-pph-.pp^PP^ C/!) P^ ddd-r^ddy-^d-r-^'T^'r^ddddd'r^ddddd-r^-r^d'r^ CN t^ CM eg O CO CO o> 00 CM CO oo r^ inooooincgcococM"^!^ U OOCOCOCOOOI**r^CMOCOCOCOCMCMOI^inOOI^CMCgCMCMOC» -a CO CMcgincooocomT-co^co CMr^or^cDegcMcgT-coT-cocoinco CJ>OCJ>r*-^OCMC»00CD"* ^CMCMT--»-i^cg"*CMOininincDCD ^ d T-" d -r-' t-: T-: T-: d T^ "r^ T-^ p/ D •Uol uo (^ CU ,^^ CJ>OOCOOOOC3>CMCJ>T-CMC3)^^T-COh>-OegcOI>^COCOinCD-«-^ inc3>ooooo>oO'«-h-cDocooocDC3>h«-coT-coo)h*^inr«-co^ '5? CO 0Or^^CMCM0OCOinC»CO0OC3>0OCM^CMr*-^0O"*COh- IM c*^ u ooc3)oo"*"^cocOT-inininT-h»r«-.cocooor>-incj>oooocoor^-«- i^inegooh-.incgT-cDcDCMCM^ocoocj>inoooT-oooeg-«- 00 c3>'0"h-cOT-cocoT-h*inoin^in'««-incocor^r>-CMco^CM^cD a> >. h-cj>cj>coocj>egoo^egr^-«-cj>cgcj>oin^^c»in^inT-egT- J t3 OOtD-r-f-Of-Oy-T-t-r-O O -r- r- r-OO-r-r-T-y-T- JS CO ^COI^I^O>COOh-CMCO^CDCDin'^T-^COOI^r^O^WT-B00 T- o00 CO incoc3>'0"h-r*-^cDi^t»-ooincocooooooocoegcoin^"^C3>'0- 00 3 C3>CMcoegT-coincgT-coor^cj>T-r^^ooc»cocj)oooooocDcoin &> T3 in CO ^ 126 ^ r«- "^ CM ^ in h-in^coc^ooo-^O'tcor-'^oocO'^cocDCM •«- CO CO CO m ^ fin-.-CRoocMoooo'^'^co^oicor^OT-flocD;* ]J eg ^ o oo r^cor^CMCDr--^ino>r*-''-oinT-coininoo>coc M o> eg CO eg o CD r^CMoor^mcomoor^incoppi^ppppf^P J2 X o o f- o dddddddddddoO'T-y-y-'^- ^•*- CMCOOO^OOT-mcOOOO^^COCMCO^OOOOOOI^OCOCDin^ ^oocMcoinoococoin"^cgc3>^T-oooc3>cj>cg^incoinir 00 "P cocoi^coooc3>cj>cM'*cO"*h-r^i^'0-o>r>-cMCMr^coocor^5to 4> C/3 CaU> p^T-;pp^r-oqph-;in^^h-;r-;p^ppppi^ppp'7 T^T^r^d'T^dddddddddddddd-r^ddy-'r-d-r- r^ i^coinT-cDcooocMOcooococo^r^co^coocMooo>int-r^ oa>a>cocMa>^ocDO>oooocoincDO>CMr^cgooo>^^ooh-^ ^ 1? CMCMcot»-cDinco^in-»-ooincj>cgc»OT-co^cocj>f>^cDCMOO pcM^^ppi^pr^pinin^r^cD"^pppoipppcMpp c/D cu y^r-^T^y^dddy-^ddddddddddd-r^dd-r^-r^dd CN r>^cocoo>T-incor«-CMO>^c»cDincDoor»-T-r^in^cocOT-^'^ TG O cooooocMC»r^oocDcgcoooooego^r^inococj>T-cDOCJ>cor«- X CO r>-h-coinoococoin'*coinoooocg^int^cocoegcj>cooocgocg ^coooinQO'«-ino'r-r^oot^cocMoooooa>-^ooega>cooooo 1$ ^T-^cMCMT^d^-r^T-^T^ddd^^dd-r^ddd^i-^csi^T-: "^ CX -o p (^ cu r^ooin-«ooooooo^c3>cginooi^-<-in^cDooocooo^r^-r-co inincgrocoooD>cDinocDh>cocM^co^incM^cococ3>cocgin cDh-CMooh>-oocooinini^cMoooocj>int-cor^r^cocoooi^cj>o ^oooo>inr>-oc3>c3>T-inoor^coT-ooc3>a>oocDC7>coo^r^o T-T-eg-f-^Ot-T-OT-oooT-^ooooooT-cMCM CM CMoocMoocMoocor«-coo)egcj>cgcooT-T-ocMcoT-inh>.cMC3>"^ 00 _ incOCMCO-«-cOCM^Cg^CJ>r»-"*C3>O^CDCO"*CJ>C3>00"*OOC3>l^ cOT-oooT-ino>ocD'»-r^^r^ooin^^r^inco^T-cocDooo Oh>i^r^inc3>ocDOC}>coocoooaoh»r^c3>c}>oooini^'^in JS ^ ^T-'^'^p-'r-^d^'T^-r.^ddT-'dT^T-^dddddd^'r^^T^^ ir> •"• r^CM0)0)T-'^-t^0>00C0000>O^C0OT— 0>^00'^000>'T— ^o> Vi •»-co^T-oh«-i^ooh-cocoi^coinoocMinoh-ot^T-^-coc3>o 00 3 cocMcoooinr^coino>co^^oO'r-cj>r«>.ino)o>inT-i>^^cocM CMOCJ)r^Cs|ppp'^CM"^h-h»-CMpCMpT-;pr^pt-<3>CMr^"^ J O cdin-^^-^'T-^CMcdcMCM^-r-^-r-'cdcsicM-r^oicMCsicMcd'^cd"^-^ >-» p ir» ^ CU ^T-in"*oo'*oooc3>T-c3>egcD'*ooi^O"^r«- •«- 00 c» CO T- eg ooeg^cMr*-T-egincgT-r^inegincoo>inocoin ^ 00 CM CO CO o 00 CO COCMCOOOOOOCO'^-mincOCMCMT-OOOOh-CMO ^ O T- CM 00 CO r-CMCMOOCDOOp'«-;ppppppppCMpp 3 « 3 ^ p p p ^ eg _C •csicMT-^^'r^^^'t-^ddd^T-^dd^^'r^ ^ -r-^ CM cd Csi Csi xi CO G »;: CJ O CJ oincgoocOT-cj>T-incoinr>-inocO"*C3>or*.ocDoocj>"^co CMCOOOCJ>CO'*CDOCOCM^COCDr-CMC»COh*CO-«-T-T-^00'^00 ocoi^^cO"*or-^egr-oc3>^"*h-cooocDO^cocj)C3>o>o HJ 2 pp'^'r^cDr^ocMa>pcMh-_pp^ppppr^ppr^t^incM CMcd-^'^'cdT-^csicd^cM-r-'-r^^cMCsiT-^^T-'^cg-r^cM^'in'^'* 1 127 o>OT--,-cMoegcM<»cMoooooo r^ ^ 00 eg oiooT-ino^CMomT-oooooo to O Oi Oi ojr^or^CMcoooco-^oocDcoincDin 00 CO •«- CO o ^ pph<.ppooo>ocM'«-oor^ooppr^ 00 CM p p ja X dddddddy^y-^y-^dddddd d d d d cMh-cooococomr^ocooooooo CO to Oi CO cDinr--cDT-cor^coincooooooo c JS to Oi r^ -^ o to CO t^ Oi ^oocMcor^oococomooooooo CM oo D> oo 00 •a r<-^cocoooc»oir>-incj>oooooo ^ m •«- CO V o oocMoor^coa>cDCMOcoa>cDcoT-oa> 00 CM O O C/D PH PpppppCD00C3>'<-C00000OC}>00 dddddddddy^dddy-^dd d d d d 3r d An t CJ>0>CM"*00h-T--^CMCDOOOOOO 00 00 ^ CO i^ooi^^'«-oooocor*-cooooooo 00 T3 in^f-0)incM^cooo"*cocO"^cOT-cM CJ) r^ r^ -^ V ppppincor^h-i^ococOT-T-T-T- C/3 CaU> O CM T- CO dddddddddddd-r^y^y-^T^ 00 CM o o 2r d An t d d d d inCDCDCMCOCOC3>OOCOCMOOOOOO Oi Oi to o o cocMCDegcMcocDin"^c3>^ooooo tn ifi yt Oi xil i CO co^incoooooincM'*o>"^c»ooi^i»-CM eg CM CM ^ CO cD^o-^^cOT-floincD-f-cMCDincoh- CO CO o o T-' d d d p/ D •o o bs u m CU ^ oocoo>r^oooor^^c7>ooooooo o CO 00 1^ cocor>-cginincMOD>o>oooooo ro- o a> 00 "S CO o)ooegc3>coini^r^cocMCMCM^co^^ a> C7> CM in CO 3 in^ocoegcO'«-r-cj>h-ooincoot^o ^ CO o o -r^ d d d S ar p T3 mu u 0)CDinegoooh-"*cj>"^o o o o Oi Ol CO Oi oa>cor<>r^a>r^coino>o o o o 00 CO CO in 00 cDincoincMO>oocDoooir« CO 00 CM •* 4> ">» egoocDO>T-oor-^cocor»- rC^O CoO CO CoO HJ o o o t5 T-^ d d d ^' d d -r^ T-* T-" d T- T- T- T- o o o JS CO Q inocoococomininT-oooooo O O •«- CO CO h- C» CO ^ 00 3 00>C3>C3>I^COCM^a>0000000 CO in t^ •^ cDOO>ooin^c3)^cj>T-in^coincocD a> ts cM^T-r^coppppr^'^^ppesjp "^ p p •«- J O CO d d d JS CX cdoicMCsicgT-'oicd-^cd'^'^iri'^'inin m Uol Pu r^coooT-^o^r^coooooooo r^ o CO r^ T-cMcoco^r^ooegT-cMoooooo m CO in o> •a 00 Vi coo>incj>r^cocMOC3>cooT-inincg"^ T- CD CO CD 3 h-'t-CMcocoininoooooooocooocoinco a3> a> 1^ -* o o C J B- ^^'•r^T-^^T-^T-^T-^r-^T^esiCMOicgoiCM y^ d d d ^^ JS CO G ir» CJ O U o>inr«-coocooT-c3>cooooooo CM O CO CO ^ CO T- cj> r^ 00 Vi OC3>OOC0^CDC3>in-<-OOOOOO m 00 m T- CMcoT-a>oooegincMcoo>h>inc3>o>oo C3> h- o T- < pppppc7>pcop^pegh^pcMCM 2 csi d d d co'oiesiesicMCsicsicdcdcdcdcd"^cd"^^ abl e H 128 00 CO O) T- m o> in ^ocj>oc»cO'*oo^^ooinoocO"*o^ •* cj> h- "«- h^ ^.inr^oooT--«-CMOOO)CD-«-i^T-ococM ^ '(-t^r>.h.cococooor«-in^oocDcococ7>cocO'^CMCDoocoinoco -* o O H CMT-coincooocoooinooi^cocococDCMCMcoin^cocD^^i^in .2 X y-^y^y^ddddddddddddddddddddddd OitOtOOOtny-y-tOt^'^COOiO ooinincO'^r^cj>egf-t-oo COt^Oiy-yr-y-OtOtOy-Ot^Oiy- cgcj)0"*T-c3>cocoegoor>-r^cococM'*cj>h*-i^oocM'«CMCM"^"*CMCO"«-T-cOO-«--«- - o^ S ooco'O'OOoocoooooT-cMCMOcocoinor^r^coinr*- OCD^CO •^ X CM eg Ol cocoegcMcgcgcMcocM in 1^ t^ CD OQ ^i^cj>cgin'0-r«-coc3>^ino^oegt>«-cD^^'<-r^in'^inT-oo oo-«--r-T-^cgoh-T-inc3>CMcoh-CMT-in"*"*ooinc3>r^oinoo T3 inocgT-ininr^int-^cDOooh-ococoi^cMooc3>cD-«-c3>oo a> O ^ cROT-cor^r^h-mcRooppp-t-h-poipp^pco^coinin 3 ^ X dy-^y^dddddddy^ddy^ddy^y^dddddddd ^C 0) «*H --••J G O c»^ooo-«-T-egcDoo-t-ino)h«-ooco^incoh*in^h-h-oor^co inegocMcoi^o^cocDr>-i^o^cDoo^^egco 129 oococMincoo>cj>co CMoegoino>inc3>OT-cMooc3>egT-T-inco^c3>incD-r-oocDCM cocj)CMr-cj>^coooincoocMegi>>-cocoi^cMCj>^inoooooinco 51 i^ooco-<-coocoincD^i^ooooc7>'<-coa>inooa>-<--<-oegoo^ c (i^ pinh^oocoh-cj>"*T-ooino)oo^cM'r-;hopcDCD el s X H ffi •«- 130 r»-r*-oocj>t-^egoocDT-cot-cj>cj>^ini^cocMcoooo-«-ocD ooi^coooT-cR^^'^cocococooo'^cD'^r^cOT-h-.'^oooocoin a>cor^cMegcDina>oco^-^coa>coc3>cot^cMco^a>oco^oo O H ^CslCMOIO|pppCOincOCD' OOOCMCMCMO cgr«>-»ncor^incgoocooooocor«-oo^c3>co^oor«-coo>co'^CMoo 0"*oococ3>inT-cg^<»cj>h-incD^^r«-oo"*T-egcj>ooooooco T3 oocoo>o>coh-ooT-cgoo)o>inocg'«-o>T-cooo'^oocoocoo> o •"• inincocDininooin^in^eg^r^oocj>co^"^cococoincococo g jn X dddddddddddddddddddddddddd o CJ co^coinor^cMoor^^incginegcDOO-«-'^"^oo^^T-T-i^ co'*c3>CMCDCM 131 incMcocM^or>-eoincDCMco^oooco^inincMegr^cMoocj>co o^r^coinoo^T-o)oocoo>oocO'^o>T-r«-T-o>"«j-cj>o>coo>o G r. r^oegooooco^^incoc3>CMr^cj>coT-r^-«-T-c3>OT-r>*ocoin o H pcsiph-;ppcsippcM^T-;esjpcMpCMCMT-T-iocgcoegocM i2 X dy^ddddy^ddy^dddddddddddddddd cocMcooincMincM"*i^oo^^coocMinc3> ^oo"*ooor«-h-CMinh-h-r^inor*-h-h«. egc3>^C3>OT-cooocj)OinegT-inoo^f^cMOoocor-cocM^^ •t-oooh-T- 132 CMinT-eg^cj>^"*h^ocM-«-cj>cDininh-egr^c3>oooorocoini^ OOt^T— 0>OC000C0O00'^-00^-h~'*— ^— ^coooh~ooo>oinoo cD-«-eg^incocMco^ocMeg^OT-h-incoincMCj>CMCJ>^"*i^ O H CMeg^co^^cMCMCMcgcMCMCMCMegeg-<-cocM^-egoooocj>co .2 X ddddddddddddddddddddddy^yr^dd ^oinoo-«-T-cMi^r«-.CMCMegcgr«*h-h-^-inT-cgr«-ooo"*-«-o ooo^cocDCMO^inin'«-cj>co<»cDininT-cocj>cgT-ocMooco c cor^dcDOooooocMO^— Oh-**"*'"''*"^'*''^'"'"''^"^'"'"^ o •^ininr^ooc3>cD"t-cj>oo^'«-co TtCDOC7>inO>CMOOOCMPOin S •^^ *ti ttt r«- uw WJ «w ^^ w «w ^^ ^^ t'* -^^ OT-ocoegc3>ooc7>ocDCJ>c»-.»•»—»—»»"* *>i wj «j w* «-j «.^ w* MI>i ^•r^^^cd^^CMoiesiesicdcMCsicsioicMcdoicM-«-^T-^f-^csi^^cM el s X H PQ inoooT-r^-incocoi^cocDoor-^ininh-coo^inot-cMcooo cgh-.incoo^T-ocj>cocO"^"*"*"^^'«-r«-Oh-cooooooooooo c3>oa>r^in-^oocMinooooo-r-a>c3>oooo'^cDcoo^^«o-oo X ^ppph>>pppppp^ppph>.CJ>h>;OOCOCDC3>h>C3>C3>C3> nue d Iso n '§ S ddddddddy^y^dddddddddddddddd o u pin^cMcoco^T-c7>in-^oo-i-oocMooinr^co^^cooo^cDco inr^oo'<«hfloinincoocoo^inin^in^^eg^c3>cMcDr^coc7> co^-cDinco^cDinoo>cocoooo^-^o>oincDCMO>ooooa> '^p^r*»p^h-.pp©r^^pinincocococDinini*^0)r^r*-co <• so n X oooodddddddddddddddddddddd 3 J CO H 133 co^cococO'^egT-^o^cor>."*r>.incj>ooincM-<-T-inincoco T-cocoococMco^^cocor>.oooocoooT-eginr^cMcocoor>. or>-coinooegcMoocgcDcOT-'»-ocoh*r^eMoocj)0>'*r^cDinco O H oocj>OT-ooin' oegc»oco-«-^oooin^in'*cDooooocMC3>' r>-^T-cMinc3>cj>inincj)C3>cDCJ)inr«-o>h-cocococDinr^c3)^'«- cj>cooincDin^o^incMooocj>h-cO"*oco^oococor*.c3>-«- cooooo-»-r^cj>r^ooco^CMco^T-ooegcMCDCD^o-«-o>or^"* h«;CMCMh-^p^pph-i*^in^incoin"*"«*CM^coincD-*-^'»--^ X Iso n nue d dy^y^ddddy^ddddddddddddddy^-r^y-^y-^ S H S o U co'^CM^incMoo'*ooc3>eg-»-coo-«-"^coco^ocMT-inT-cocD ^"^^Oi-r-OiiO'^COCOy-Oaiy-toaiCOOiOitOy-y-OOiy-tO I"2SS9S2*^'^^*>^<'>''><^<*> 134 __:.iiHr- Id 3 < CJ H CO 4> O G nued H H CJ Vi O G Vi G O son CJ Ison H X :^ NJ PQ HJ S X X X '^opq>inincga^a>a>cooooooo oocMoococooincocomoooooo cMeginoocMCJ>ocDO^-«-CMoocMr>-eg oh^coooomoocoh^cDOOOOOO -^OT-T-cocMino-^^oooooo OOOOOOOT-OOOOOOOO "«0-00 GENUS Palaemonetes Heller 1869 FROM NORTH AMERICA (CRUSTACEA: DECAPODA) JAMES THOMAS COLLINS, B.B.A., M.S A DISSERTATION SOI T^ ACKNOWLEDGMENTS i \ an I would like to express my sincere gratitude to Dr. Marilyn A. Houck. As r ri 0 chair of my committee, she provided guidance, encouragement, and enthusiasm ' throughout the course of my work in her lab. And also to Dr. Houck, a very special thank you for the long hours spent helping me in the preparation of this manuscript. I would like to thank Dr. Ned E. Strenth for introducing me to the genus Palaemonetes, and for his help and guidance over the years as I struggled to understand this group of decapods. To Dr. Richard E. Strauss, thank you for the computer programs necessary for the analysis of my data. I also wish to thank Dr. Llewellyn D. Densmore EQ and Dr. Michael R. Willig, both of whom provided support and guidance through the course of this study. I would like to express my appreciation to N. E. Strenth, H. L. McCutchen, and S. Jasper for the loan of specimens from their personal collections. I would also like to thank R. Manning for arranging for the visit to the Smithsonian and sending the specimens to Texas Tech University. I thank G. Longley for permission to collect from the artesian well on campus at Southwest Texas State University. The Biology Department provided me with financial support, as a teaching assistant and a summer mini grant, for which I am extremely grateful. I would also like to acknowledge Drs. Clifford B. Fedler and Nick C. Parker and Marilyn A. Houck for allowing me the opportunity to work as a research assistant under a Department of Interior grant. It is a pleasure to acknowledge the many friends and graduate students, both in the lab and in the Department of Biology, who offered their support. A special thank you goes to Sara, Doug, Leslie, Darin, and Elizabeth who over the years have listened and offered support when I needed it. A special thank you to my wife, Gloria, who lovingly endured the years of graduate school and was instrumental in my success. I would also like to express my thanks to our son. Cliff, for encouragement and fiiendship along the way. And to the ii memory of my Mother, for her support and encouragement to go forward and achieve my goal I dedicate this dissertation. Ul TABLE OF CONTENTS ACKNOWLEDGMENTS ii LIST OF TABLES v LIST OF FIGURES vi CHAPTER L INTRODUCTION 1 Taxonomic and Systematic Review 1 Research Objectives 10 II. MATERIALS AND METHODS 12 Specimens 12 Equipment and Handling Procedures 16 Statistical Protocols 17 Statistical Procedures 19 General procedures 19 Influence of gender on character discrimination 21 Comparison of species within the genus Palaemonetes 22 Comparison within species of P. kadiakensis 22 Evaluation of character contribution to discrimination 22 m. RESULTS AND DISCUSSION 223 Influence of Gender on Character Discrimination 223 Analysis of All Taxa 223 Comparison of Species within the Genus Palaemonetes 47 Comparison of Discrete Characters 49 Comparison within Species of P. kadiakensis 68 IV. CONCLUSIONS 91 LITERATURE CITED 94 APPENDED: COLLECTION NUMBERS AND RAW DATA 100 IV LIST OF TABLES 2.1 Populations collection data 13 2.2 List of characters used in the analyses 18 3.1 Rao's V hierarchy for the 26 characters used in the analyses of all taxa 40 3.2 Rao's V hierarchy for the 30 characters used in the analyses of Palaemonetes 62 3.3 Rao's V hierarchy for the 30 characters used in the analyses of Palaemonetes kadiakensis 87 A. 1 Collection numbers and raw data utilized in this study ofthe North American Palaemonetes 100 LIST OF HGURES 3.1 Convex hulls and centroids indicating the discrimination among males and females, for all taxa studied 24 3.2 Relationship between the first two axes of discrimination as determined by principal component analysis (PCA) 25 3.3 Centroids and one standard deviation about the centroids as determined by principal component analysis (PCA) for all taxa studied 26 3.4 Convex hulls and centroids indicating discrimination due to the two major shape axes as determined by principal component analysis (PCA) for all taxa studied 28 3.5 Centroids and one standard deviation about the centroids due to the two major shape axes, as determined by principal component analysis (PCA), for all taxa studied 29 3.6 Vector plot showing character correlations contributing to character variation resulting in shape discrimination, as determined by principal con^nent analysis (PCA), for species of all taxa studied 30 3.7 Vector plot showing character correlations representing major character variation contributing to discrimination, as determined by principal component analysis (PCA), for all taxa studied 31 3.8 Plot ofthe major size axis versus the major shape axis, as determined by discriminant fimction analysis (DFA) for all taxa studied 32 3.9 Centroids and one standard deviation about the centroids as determined by discriminant function analysis (DFA) for all taxa studied 34 VI 3.10 Plot ofthe major size axes, as determined by discriminant function analysis (DFA) for all taxa studied 35 3.11 Vector plot showing character correlations contributing to character variation resulting in discrimination, as determined by discriminant function analysis (DFA), for all taxa studied 36 3.12 Vector plot showing character correlations contributing to character variation resulting in shape discrimination, as determined by discriminant function (DFA), for all taxa studied 37 3.13 Plot ofthe cumulative Rao's V values for all taxa studied 38 3.14 Cumulative Rao's V values for all taxa studied 39 3.15 Plot ofthe major axes, as determined by size-free discriminant function analysis (DFA) for all taxa studied 41 3.16 Centroids and one standard deviation about the centroids as determined by size-free discriminant analysis (SF) for all taxa studied 42 3.17 Unrooted phenogram (UPGMA) of the relationships among all taxa studied, using size-free Mahalanobis distances 44 3.18 Unrooted neighbor joining phenogram of the relationships among all taxa studied 45 3.19 Neighbor joining phenogram ofthe relationships among all taxa studied, rooted with Leander as the outgroup 46 3.20 Convex hulls and centroids indicating the discrimination among males and females, for species oi Palaemonetes 48 3.21 Relationship between the first two axes of discrimination as determined by principal component analysis (PCA) 49 Vll 3.22 Centroids and one standard deviation about the centroids as determined by principal component analysis (PCA) for species oi Palaemonetes studied 50 3.23 Convex hulls and centroids indicating the discrimination (PCA) among species oi Palaemonetes, due to the two major shape axes 51 3.24 Vector plot showing character correlations contributing to character variation resulting in discrimination, as determined by principal component analysis (PCA), for species of Palaemonetes studied 53 3.25 Plot ofthe major size axis versus the major shape axis, as determined by discriminant function analysis for all species oi Palaemonetes 54 3.26 Vector plot showing character correlations contributing to character variation resulting in discrimination, as determined by discriminant fimction analysis (DFA), for species of Palaemonetes studied 55 3.27 Plot ofthe major size axes, as determined by discriminant function analysis (DFA) for populations oi Palaemonetes studied 56 3.28 Centroids and one standard deviation about the centroids as determined by discriminant analysis for species of Palaemonetes studied 57 3.29 Vector plot showing character correlations contributing to character variation resulting in shape discrimination, as determined by discriminant function analysis (DFA), for species oi Palaemonetes studied 58 3.30 Plot ofthe cumulative Rao's V values for all species of Palaemonetes studied 60 3.31 Cumulative Rao's V values for all species oi Palaemonetes studied 61 Vlll 3.32 Plot ofthe major axes, as determined by size-free discriminant function analysis (SF) for all species oi Palaemonetes studied 63 3.33 Centroids and one standard deviation about the centroids for the shape axes, as determined by size-free discriminant analysis (SF), for species of Palaemonetes studied 64 3.34 Unrooted neighbor joining phenogram ofthe relationships among species oi Palaemonetes studied 65 3.35 Unrooted phenogram (UPGMA) ofthe relationships among Palaemonetes species studied, using size free Mahalanobis distances 66 3.36 Convex hulls and centroids indicating the discrimination among males and females, for populations of P. kadiakensis 69 3.37 Relationship between the firsttw o axes of discrimination as determined by principal component analysis (PCA); convex hulls with centroids for populations of P. kadiakensis studied 70 3.38 Centroids and one standard deviation about the centroids as determined by principal con:5X)nent analysis (PCA) for populations of P. kadiakensis studied 71 3.39 Vector plot showing character correlations contributing to character variation resulting in shape discrimination, as determined by discriminant function analysis (DFA), for species of Palaemonetes studied 73 3.40 Convex hulls and centroids indicating the discrimination among all taxa (PCA), due to the two major shape axes for P. kadiakensis populations studied 74 3.41 Centroids and one standard deviation about the centroids due to the two major shape axes, as determined by principal conqwnent analysis (PCA), for populations of P. kadiakensis studied 75 IX 3.42 Vector plot showing character correlations contributing to character variation resulting in shape discrimination, as determined by principal component analysis (PCA), for the populations of P. kadiakensis studied 76 3.43 Plot ofthe major size axis versus the major shape axis, as determined by discriminant function analysis (DFA) for populations of P. kadiakensis studied 77 3.44 Plot ofthe major size axes, as determined by discriminant function analysis (DFA) for populations of P. kadiakensis studied 78 3.45 Vector plot showing character correlations contributing to character variation resulting in shape discrimination, as determined by discriminant function analysis (DFA), for the populations of P. kadiakensis studied 79 3.46 Vector plot showing character correlations contributing to character variation resulting in shape discrimination, as determined by discriminant fimction analysis (DFA), for the populations of P. kadiakensis studied 80 3.47 Plot ofthe major axes, as determined by size-free discriminant function analysis (SF) for all populations of P. kadiakensis studied 81 3.48 Vector plot showing character correlations contributing to character variation resulting in shape discrimination, as determined by size-free discriminant analysis (SF), for the populations of P. kadiakensis studied 82 3.49 Centroids and one standard deviation about the centroids for the shape axes, as determined by size-free discriminant analysis (SF), for populations of P. kadiakensis studied 84 3.50 Cumulative Rao's V values for popiilations of P. kadiakensis studied 85 3.51 Plot ofthe cumulative Rao's V values for P. kadiakensis 86 3.52 Unrooted phenogram (UPGMA) ofthe relationships among Palaemonetes kadiakensis populations using size-free Mahalanobis distances 88 3.53 Unrooted neighbor joining phenogram ofthe relationships among populations oi Palaemonetes kadiakensis, given by county of collection 90 XI CHAPTER I INTRODUCTION Taxonomic and Svstematic Review The genus Palaemonetes Heller, 1869, was first described as Palaemon varians from the eastern Atlantic by (Leach 1814) (Holthuis, 1952). Since that time, several species oi Palaemonetes have been described from all ofthe continents with the exception of Antarctica. In North America there are currently 14 described species {Palaemonetes vulgaris (Say, 1818); P. paludosus (Gibbs, 1848); P. antrorum Benedict 1896; P. kadiakensis Rathbun, 1902; P. hiltoni Schmitt 1916; P. intermedius Holthuis, 1949; P. pugio Holthuis, 1949; P. cummingi Chace, 1954; P. suttkusi Smalley, 1964; P. lindsayi Villalobos and Hobbs, 1974; P. texarms Strenth, 1976; P. holthuisi Strenth, 1976; P. mexicanus Strenth, 1976; and P. hobbsi Strenth, 1994). Of these, 13 species occur in the eastern United States and Mexico, and only P. hiltoni is known fromth e southwest coast ofthe United States and the northwest coast of Mexico (Fig. 1.1). Although these species oi Palaemonetes are an important part ofthe temperate and tropical aquatic food webs (Lowe and Provenzano, 1990; Garcia, 1991), they have little commercial value (Strenth, 1976). In North America, their value has been restricted to use as commercial fish food (Worth, 1908; Nielson and Reynolds, 1977); however, in Viet Nam, P. camranhi, Palaemon semmelinkii and Periclimenes grandis are used as a dietary supplement for humans (Xuan, 1997). The lack of commercial importance, the occurrence of morphological homogeneity among taxa (Strenth, 1976), and the lack of "understanding ofthe currently used taxonomic characters" (Strenth, 1976, p. 1) contribute to the lack of resolution ofthe systematic relationships for this group of natant decapods. Two ofthe freshwaterepigea n species, P. paludosus and P. kadiakensis, range over a wide area. Palaemonetes paludosus is found primarily along the east coast ofthe United States from Florida to New Jersey east ofthe Allegheny G G Gu ^ E G B s: "S o s: ins B «^ -Q S CO CD •1^ • i-H CO o -«-) CJ 'o o ^ CM O o »-( CO 0 4:3 • l-l Mountains (Holthuis, 1952). However, several authors have reported iX)pulations of P. paludosus ranging across the southern United States as far west as Texas (Creaser and Ortenburger, 1933; Meehean, 1936; Holthuis, 1952; Webb, 1980; Garcia, 1991). Palaemonetes kadiakensis ranges from northeast Mexico, throughout the Mississippi River Valley, and into Wisconsin and Michigan (Creaser, 1932; Nelson, 1982; Hobbs and Jass, 1988). The other freshwater epigean species exhibit restricted habitats. Palaemonetes lindsayi, P. suttkusi, P. mexicanus, P. hobbsi, and P. texanus, are artesian spring forms only found within a few hundred meters of spring discharges. Subterranean species also exhibit a fairly restricted distribution. Two ofthe subterranean species, P. antrorum and P. holthuisi, are known only to occur in the Edwards Aquifer of central Texas. The type locality for P. antrorum is an artesian well on the campus of Southwest Texas State University in San Marcos, Texas. However, it has been reported from the Edwards Aquifer (Ulenhuth, 1921) as far west as Uvalde County, Texas (Hobbs et al, 1977). Palaemonetes holthuisi also is found in the Edwards Aquifer. It was described from Ezell's Cave (Strenth, 1976) which is only a few miles west ofthe type locality for P. antrorum. Ulenhuth (1921) believed that both the artesian well at Southwest Texas State University and Ezell's cave are part ofthe subterranean Purgatory Creek System. Webb (1980) reported a single individual of P. holthuisi collected from the artesian well on the campus of Southwest Texas State University. Garcia (1991) also reported having found a single individual of P. holthuisi but fails to give the location. The marine or estuarine species, P. vulgaris, P. intermedius, and P. pugio, have wide geographic distributions. Holthuis (1952) in his revision ofthe family Palaemonidae, noted that all three species range along the east and south coasts of the United States from Massachusetts to Texas. Palaemonetes vulgaris generally inhabits brackish and salt water, while P. pugio occurs in brackish to abnost freshwater (Holthuis, 1952). The single marine or estuarine species from the west coast of North America, P. hiltoni, ranges from southern California to the Mexican states of Sonora and Sinaloa. According to Holthuis (1952), little is known about the habitat of this species; however, he notes that the specimens collected from Del Mar were collected in a slough and those from Sinaloa were from an estuary. When Holthuis (1949; 1952) revised the genus Palaemonetes, his classification and taxonomic descriptions focused on the identification ofthe currently described species rather than on their evolutionary relationships. Since that time various authors have produced studies on population genetics and systematics on some species oi Palaemonetes fromNort h America (Webb, 1980; Collins, 1993; Garcia and Davis, 1994). Pereira (1989) examined the superfamily Palaemonoidea and proposed systematic relationships above the generic level. Although these studies have provided information about the taxonomy and systematics of this group, the systematic relationships among species ofthe genus Palaemonetes from North America remains unclear. Based on egg size, Sollaud (1923) divided the family Palaemoninae into two groups, one with large numbers of small eggs and the other with fewer numbers of large eggs. During an examination ofthe different species listed in the revision by Holthuis (1950), Strenth (1976) found that marine species exhibited numerous small eggs and freshwaterspecie s exhibited fewer but larger eggs, and concluded that egg number and size substantiate the separation oi Palaemonetes into two groups, as first proposed by Sollaud (1923). Numerous studies (Broad, 1957; Broad and Hubschman, 1962, 1963; Dobkin, 1963, 1971; Hubschman, 1974; Strenth and Longley, 1990, Strenth, 1991) have documented developmental differences among taxa based on the number of larval stages. Freshwater species with larger eggs generally have fewer larval stages whereas marine forms exhibit more larval stages. Palaemonetes kadiakensis is the exception to this rule in that its development more closely resembles that of marine taxa (Broad and Hubschman, 1963). Generally, the freshwateran d the marine groups can be separated on qualitative morphological criteria. Holthuis (1949, 1952) noted that there was a difference in the number of fused segments and the length ofthe unflised distal end ofthe upper antennular flagellum. Freshwater species, with the exception of P. antrorum and P. holthuisi (Strenth, 1976) have the free part of their antennule shorter than the fused part. In marine species and the freshwaterspecie s P. antrorum and P. holthuisi the free distal end is longer than the fused portion (Holthuis, 1949, 1952; Villalobos et al., 1974; Strenth, 1976). This similarity between antennule morphology might be attributed to an adaptation to low light levels by the subterranean species that are tactility oriented and have degenerate eyes. Generally, a difference exists in the antenna morphology ofthe form I zoea stage of freshwater and marine Palaemonetes. With one exception, P. kadiakensis, the North American freshwaterspecie s oi Palaemonetes exhibit an unsegmented antenna scale, whereas marine species have a segmented antenna scale in their form I zoea. Differences in egg number and size, antennule morphology, and development between freshwater and marine groups, strongly supports the monophyly ofthe freshwater species oi Palaemonetes fromNort h America. The evidence examined by Strenth (1976) led him to conclude that the freshwaterspecie s oi Palaemonetes from North America are in fact a monophyletic group. Since Holthuis' work (1949, 1952), there have been several attempts to explain the systematic relationships within Palaemonetes. In his revision ofthe species oi Palaemonetes from North America, Holthuis (1952) placed P. antrorum in the subgenus Alaocaris. AW remaining North American species were placed in the subgenus Palaemonetes. This separation was based on four characteristics: (1) degenerated eyes without pigment, (2) absence of teeth on the lower margin ofthe rostrum, (3) similarity ofthe shape and size ofthe first and second pereiopods, and (4) lack of a movable exopod spine on the uropods. With the description of several new species oi Palaemonetes from North America, the separation of P. antroum into its own subgenus became untenable and Strenth (1976) synonymized the subgenera Palaemonetes and Alaocaris. Vandel (1965) noted that it is common for eyes in cavenerincole organisms to be reduced. This is especially true of subterranean species oi Palaemonetes from North America (Benedict, 1896; Chace, 1954, Strenth, 1976). Strenth (1976) noted that both P. antrorum and P. holthuisi did not have teeth on the lower margin of their rostrum. The reduction of lower margin dentition is common in the subterranean species of palaemonids (Strenth, 1976). There appears to be a correlation between the reduction of lower rostral dentition and degeneration of eyes, which Strenth (1976) attributes to a "specialization associated with life in the subterranean environment" (p. 14). The shape and size ofthe legs of P. antrorum are also similar to those of P. holthuisi. These two species feed while suspended on their last three legs, using the first two pairs of pereiopods extended, rather than the crouched position typical of epigean species. Leg morphology is an adaptation for living in a subterranean environment (Strenth, 1976). At the time Holthuis (1952) revised the genus Palaemonetes there were only three freshwaterspecie s described from Nortli America. Palaemonetes antrorum differed from P. paludosus and P. kadiakensis in that it lacked the movable exopod spine and was from a subterranean environment. Since this revision three additional freshwater species oi Palaemonetes have been described that lack the moveable exopod spine (Smalley, 1964; Villalobos and Hobbs, 1974, Strenth, 1976). With the addition ofthe epigean species, which do not exhibit a movable exopod spine the validity of this character for the separation of P. antrorum into a separate subgenus is greatly reduced. Although the variability ofthe movable exopod is quite high, Strenth (1976) stated that it is an important character for understanding the systematic relationships among the freshwaterspecie s of this group. Based on the condition ofthe movable exopod spine and geographic distribution, he divided the freshwaterspecie s into three groups. The first, consisting of P. kadiakensis, P. paludosus, P. cummingi, exhibits a movable spine on both exopods and is distributed from north eastern Mexico to the east coast ofthe United States. Palaemonetes antrorum, P. holthuisi, P. suttkusi, and P. lindsayi compose the second group, in which the movable spine is absent on both exopods. This group is distributed from central Texas to Mexico, west ofthe Sierra Madre Oriental Mountains. The range ofthe third group extends from central Texas into Mexico along the eastern side ofthe Sierra Madre Oriental Mountains. In this group, the movable exopod spine "is quite variable; both movable spines may be present, both may be absent, or only a left or right one may be present within a single population" (Strenth, 1976, p. 5). At the time Strenth divided the freshwater species into these three groups, this third group consisted of P. texarms and P. mexicamts. Subsequently, P. hobbsi, was named (Strenth, 1994). This species is morphologically consistent as to the movable exopod spine. In addition, it is a geographically intermediate between P. texarms and P. mexicanus. The relationship among the three groups was examined most recently by Collins (1993) using allozyme electrophoresis. This study was restricted in scope to the species from Texas and Mexico. However, there was strong evidence to support P. texanus, P. hobbsi, and P. mexicanus as a single group. The study also indicated that P. lindsayi, P. suttkusi, and P. antrorum may indeed form a natural group. Because only P. kadiakensis was examined from the group containing spines on both exopods, any possible relationship among the member species of this group could not be addressed. It should be noted; however, that P. kadiakensis was not included in either group. Further studies are needed to determine if P. kadiakensis, P. paludosus, and P. cummingi form a monophyletic group (Collins, 1993). The use of male external genitalia as a taxonomic character was proposed by Fleming (1969) in the hope that they might resolve systematic relationships within epigean species oi Palaemonetes from North America. Using apical and subapical setae from the tip ofthe appendix masculina, he found that P. kadiakensis, P. paludosus, and P. pugio can be discriminated from each other and that they can be distinguished from P. vulgaris and P. intermedius. Since Fleming's work, additional species have been described and male genitalia subsequently have been used in the description (Villalobos and Hobbs, 1974; Strenth, 1976, 1994). A study by McCutchen (1983) on the morphological variation ofthe North American freshwater species oi Palaemonetes found features ofthe appendix masculina did separate marine and freshwater species. However his results clearly indicate that within freshwater species the variation of spines within species is greater than among species. McCutchen (1983) concluded that the appendix masculina was not a valid taxonomic character for the freshwaterspecie s oi Palaemonetes from North America. Until Holthuis (1949, 1950, 1952) revised the family Palaemonetes, P. paludosus was considered to be conspecific with the Palaemonetes from the west slope ofthe Alleghenies (Dobkin, 1963). However, as was pointed out above P. kadiakensis exhibits valid characteristics which discriminate it fromP . paludosus (form I zoea morphology and number of larval stages). According to Dobkin (1971), even though these differences in larval morphology and development occur, the adult forms appear to be more closely related to each other than to other species within this genus. Dobkin (1963, 1971) proposed that P. kadiakensis may be a recent immigrant into the freshwater environment. He has noted that with the exception of P. argeninus (Nobili 1901) and P. kadiakensis, the known freshwater species of Palaemonetes exhibit an abbreviated larval development. This he attributes to the possibility that species that have been in the freshwater longer would tend toward the abbreviated larval development. Based on this, he hypothesized that P. kadiakensis 8 may be a relatively recent immigrant into the freshwater environment (Dobkin. 1963. 1971). Palaemonetes argentinus exhibits characteristics more closely related to the marine species which lead Strenth (1976) to conclude that P. argentinus probably is related more closely to the marine species, and represents a different evolutionary pathway into freshwater at least in comparison with the freshwater species from North America. Based on the larval development, form I zoea antennular morphology, and similarity among the early larval stages of P. kadiakensis. P. pugio, P. vulgaris, and P. intermedius the hypothesis of Dobkin (1965, 1971) and Strenth (1991) that P. kadiakensis represents a more recent invasion into the freshwater environment is supported. An examination of allozymes and DNA ofthe Palaemonetes of Texas revealed two different haplotypes of P. kadiakensis based on genetic disequilibrium of selected markers (Garcia, 1991; Garcia and Davis, 1994). Populations of both haplotypes were sympatric in many locations and maintained reproductive isolation. However, hybridization occurred in several locations in central Texas (Garcia, 1991; Garcia and Davis, 1994). Garcia (1991) erected a clade ofthe Texas species of Palaemonetes based on 17 loci, {P. paludosus, P. texanus, and P. holthuisi). Palaemonetes kadiakensis was the sister group to this (Garcia, 1991). Webb (1980) re-erected the subgenerus ^/aocam to include P. antrorum and P. holthuisi based on the fact that both species live in a subterranean habitat. His electrophoretic data suggested that the Palaemonetes from Texas should be divided into three groups; (1) a marine group consisting of P. pugio, P. intermedins, and P. vulgaris; (2) a freshwatergrou p (P. texanus, P. kadiakensis, P. paludosus); (3) the troglobites, P. antrorum and P. holthuisi. Separation ofthe marine and freshwater species does not appear to be in question (Strenth, 1976); the separation ofthe freshwater species from the trogolobites is not accepted by current researchers of this group. In a re-evaluation of P. holthuisi, Bruce (1993) elevated this species to the new genus Calathaemon. He based this on the fact that the mouthparts are radicallv different from those of other Palaemonetes. In Palaemonetes the mouthparts are adapted primarily for scavenging and possibly predation. The mouthparts in Calathaemon are adapted as a filtering system (Bruce, 1993). Bruce (1991) noted that the branchiostegal region forms a cavity allowing movement ofthe caridean lobe and the scaphognathite. These movements may produce strong currents capable of drawing food particles across the filtering mechanisms. Bruce (1991) notes that this process has not actually been observed. Strenth (1976, pers. comm.) indicated the feeding habits of both P. antrorum and P. holthuisi are quite similar and not unlike that ofthe epigean species oi Palaemonetes. Bruce (1991) reported that the chelae on both the first and second pereiopods was distinctively different from that of other members ofthe family Palaemonidae. The central region of both chelae appear flattened or possible convex in Calathaemon, rather than concave as in other species oi Palaemonetes. He also noted that the chelae on both pairs of pereiopods oi Calathaemon are similar rather than distinct as in species oi Palaemonetes. Research Objectives The broad research objective of this study is to understand the systematic relationships ofthe species within the genus Palaemonetes from North America. This will provide a framework to increase our knowledge of this important part of the aquatic food chain, as well as provide a mechanism for understanding more about the biogeography and the ecology of North America. The specific issues to be examined in this study are; (1) the development of hypotheses of systematic relationships among species oi Palaemonetes from North America; (2) whether the freshwater species do form a monophyletic group as proposed by Sollaud (1923) and Strenth (1976); (3) whether following the hypothesis by Holthuis (1952), and re-enacted by Webb (1980), P. antrorum should be elevated to a subgenus (Alaocaris) and separated from the other species of North American Palaemonetes; (4) whether P. holthuisi should be elevated to a new genus. 10 Calathaemon, as proposed by Bruce (1993) is supported by the data; (5) whether the hypothesis by Dobkin (1965, 1971) that P. kadiakensis is a recent immigrant into the freshwater environment is upheld; (6) whether the origin ofthe North American species of Palaemonetes was in salt of freshwater; and (7) what the potential role of biogeography in the evolution ofthe genus Palaemonetes. 11 CHAPTER II MATERIALS AND METHODS Specimens The focus of this research was the discrimination among species of Palaemonetes from North America. However, to put this taxon in perspective relative to other genera within the family Palaemonidae, additional taxa were referenced to determine the appropriate outgroup comparison. The genera considered to be potential out-group taxa were; Leander, Macrobrachium, Palaemon and Calathaemon. Specimens which represented the genus Leander (L. tenuicornis) were collected in Galveston, Texas, and were borrowed for this study from N. E. Strenth of Angelo State University. Macrobrachium (M acanthums) representatives were collected from Montego Bay, Jamaica and Palaemon (P. northropi) was collected from Guadalupe Island in the Caribbean. Calathaemon (=Palaemonetes holthuisi), Macrobrachium and Palaemon were borrowed from the National Museum of Natural History (NMNH). Specimens oi Palaemonetes used in this research were; collected by the author, loaned by individuals that shared their personal collections, or borrowed from the NMNH. The type specimens of two species, Palaemonetes cummingi and P. holthuisi were examined on location at the NMNH. There were a total of 19 populations representing 14 species oi Palaemonetes from North America examined in this study. All nominal species of North American Palaemonetes were included in this research project. The collection locations for each species and population, number of individuals used in this study, and number of males/females are listed in Table 2.1. 12 Table 2.1. Population collection data including taxa, total number of individuals used in this study and the number of males (m) and females (f) and the collection location. Species Total No. (m/f) Location Palaemonetes P. antrorum 10(5/5) Artesian well on the campus of Southwest State University campus, San Marcos, Hays Co., Texas. This is the type locality for this species. P. cummingi 4(1/3) Squirrel Chimney, Alachua Co., Florida. This is the type locality for this species. The specimens were examined in the NMNH. P. hiltoni 9 (4/5) On loan from the NMNH. Collected in the Mexican state of Sinaloa. P. hobbsi 10(5/5) The Nacimiento de Rio Mante, Tamaulipas, Mexico. This is the type locality for this speces. P. holthuisi 2 (2/0) Ezell's Cave, Hays Co., Texas, {Calathaemon) type locality for this species. P. intermedius 8 (5/3) On loan from the NMNH. Collected from Link Port, St. Lucie Co., Florida. P. kadiakensis 9 (4/5) Sapo Lake, Hidalgo, Co., Texas 4 (2/2) Eagle Nest Canyon, Langtry, Val Verde Co., Texas 6 (1/5) Highway 36, N. of Caldwell, Burleson Co., Texas. 5 (1/4) Mullins Crossing, Tom Green Co., Texas. 13 Table 2.1. Continued Species Total No. (m/f) Location P. kadiakensis 10(6/4) San Saba River, Menard, Menard Co., Texas. 3(2/1) Colorado River at the Highway 29 bridge, Llano Co., Texas. P. lindsayi 10(5/5) 15 miles W. of Ciudad Valles, San Luis Potosi, Mexico. This is the type locality for this species. P. paludosus 10(5/5) Ashley River, W. of Charleston, South Carolina. Thought to be about 10 miles up stream from the type locality. P. pugio 9 (4/5) Goose Island State Park, Aransas Bay, Texas. P. suttkusi 10(5/5) Cuatro Cienigas Basin, Mexico. This is the type locality for this species. P. texanus 10(5/5) San Marcos River, San Marcos, Hayes Co., Texas. This is the type locality for this species. P. vulgaris 9 (5/4) Skidaway River estuary. University of Georgia Aquarium, Skidaway Island, Savannah, Georgia. Leander tenuicornis 10(0/10) Galveston, Texas. Macrobrachium acanthums 4(1/3) On loan fromth e NMNH. Collected from a brackish pond, Montego Bay, Jamaica. 14 Table 2.1. Continued Species Total No. (m/f) Location Palaemon northropi 10(5/5) On loan from the NMNH. Collections site was Caribbean Sea, Point A Pitre Guadalupe Island. 15 Epigean specimens collected by the author were captured in relatively shallow water with a hand dip net. Specimens were collected as close as possible to the type locality for each species. All were collected either from aquatic vegetation or from under overhangs ofthe bank. The subterranean species P. antromm was collected by securing a net over the artesian well discharge on the Southwest Texas State University in San Marcos, Texas. Once specimens were collected, they were placed in separate vials containing 70% ETOH. Each individual was given an identification number and the location of the collection site was recorded. The samples of populations provided by other collectors or the NMNH also were preserved in 70% ETOH at the time of their collection and permanently stored in alcohol. Individual specimens were chosen for this study based upon the quality of their physical condition. Specimens that were lacking appendages were rejected, when possible, but retained in damaged type specimens. Also, an attempt was made to select an equal number of males and females from randomly sampled populations. The NMNH collection numbers for the type species were retained and used by the author (Appendix). All others specimens from NMNH arrived at Texas Tech University as bulk-collections. Subsets of these bulk-collections were parceled into individual vials for study. The vials were then individually numbered and those numbers were retained (in addition to the bulk identification numbers) for future reference when the specimens were returned to the NMNH. Equipment and Handling Procedures Specimens were removed from the ETOH and placed in a dish containing wax so individuals could be pinned in the appropriate orientation to facilitate measurement. Specimens were measured under an Olympus SZHIO research stereo microscope with zoom capabilities l-70x. Images were displayed on a video monitor using a Sony CCD/RGB color camera connected to a Sony Trinitron (PVM1341) monitor. The images were then captured using Image-Pro Plus (V.2) 16 software from Media Cybernetics, Maryland. All ofthe specimens were measured directly from the image on the monitor. Measurements were consistently taken from the right side ofthe specimen unless that particular character was damaged (e.g., type specimens). Because the size range of individuals differed, it was necessary to image various specimens at different relative magnifications. Character measurements (Table 2.2) were then converted to millimeters, based on a cahbrated constant. The constant was calibrated by repeated measurements of a known distance at each magnification. The constant was then applied to the raw measurement. It was necessary to measure both P. cummingi and P. holthuisi on site at the NMNH. In this case, the specimens were measured using Sylvac Ultra-Call Mark III calipers manufactured by Fred V. Fowler, Co. Only linear measurements could be taken in this manner on these specimens. To test the comparability ofthe caliper measurements with others, caliper measurements also were made on individuals from the several other species, and compared to the same measurement taken from the captured image. The distance measurements were comparable (varied by no more than 2%). Statistical Protocols Anatomical landmarks and helping points {sensu Bookstein et al, 1985) were selected to define characters used for this study. Landmarks provided information about homologous points that show anatomical delineation. Helping points are non homologous points that aid in defming morphological surfaces, when recorded in association with landmarks. There initially were a total of 40 distance characters included in this study. However, the number of characters was reduced to 30 (Table 2.2) because of missing data among a subset of individuals. The Euclidean distances for the 30 characters were calculated among the landmarks for 172 individuals in order to assess character 17 Table 2.2. List of characters used in the analyses. Characters are referenced by number in the vector plots. Allometric coefficients relate to the principal component analyses of all taxa. Character # Description of Distance Characters Allometries 1. Rostrum length, from tip of rostrum to base of last 0.9664 tooth along the dorsal edge 2. Carapace Length, from tip of rostrum, along the 1.1565 dorsal margin, to the posterior edge ofthe cephalothorax 3. Ischium Leg 1, from apodeme to apodeme 1.0805 4. Merus Leg 1, from apodeme to apodeme 1.1591 5. Corpus Leg I, from apodeme to apodeme 1.0646 6. Propodus Leg I, from apodeme to apodeme 0.9861 7. Dactyl Leg 1, from apodeme to apodeme 0.8940 8. Ischium Leg II, from apodeme to apodeme 0.9392 9. Corpus Leg 11, from apodeme to apodeme 1.0974 10. Corpus Leg II. from apodeme to apodeme 0.9825 11. Propodus Leg 11, from apodeme to apodeme 0.9934 12. Dactyl Leg II, from apodeme to apodeme 0.9462 13. Ischium Leg IV, from apodeme to apodeme 1.0389 14. Merus Leg IV, from apodeme to apodeme 1.1195 15. Corpus Leg IV, from apodeme to apodeme 0.9540 16. Propodus Leg IV, from apodeme to apodeme 1.1107 17. Dactyl Leg IV, from apodeme to apodeme 1.0072 18. Ischium Leg V, from apodeme to apodeme 1.0082 19. Merus Leg V, from apodeme to apodeme 1.0920 20. Corpus Leg V, from apodeme to apodeme 0.8915 21. Propodus Leg V, from apodeme to apodeme 1.0270 22. 3rd Maxili Carpus, from apodeme to apodeme 1.0830 23. 3rd Maxili Propodus/Dactyl, from apodeme to 0.8732 apodeme 24. 2nd Segment Antenna Peduncle, from apodeme to 0.8606 apodeme 25. 3rd Segment Antenna Peduncle, from apodeme to 0.7762 apodeme 26. Telson M x M, a line aaoss the telson from margin 1.1089 to margin, located at the base ofthe anterior spines 27. Telson L x L, a line across the telson from margin to 1.0325 margin, located at the base ofthe posterior spines 28. Telson M x L, between the bilateral lines formed at 0.9972 the base ofthe anterior and posterior spines 29. Telson B X M, fromth e anterior base ofthe telson to 0.9465 the bilateral line aaoss the telson at the anterior spines 30. Telson L x T. from the bilateral line at the base ofthe 0.8069 posterior pair of spines to the tip ofthe telson 18 variation. An attempt was made to have an equal number of males and females for each species or population (see Appendix). The distances measured were chosen so as to be homologous among inidviduals and populations. Characters also were chosen on the criteria of repeatability. Only the characters that were reasonably free of distortion (due to handling) were measured. Characters were distributed across all elements ofthe animal form. Quantitative phyletic taxonomy as proposed by Kluge and Farris (1969) does utilize biological information and can be used to discover systematic and evolutionary relationships among organism. Data of a continuous nature, immunological, molecular are quantitative and must be analyzed using numerical techniques (Farris, 1972). Felsenstein (1984) noted phylogenetic distance methods are independent ofthe clustering technique used. Colless (1970) concluded that phylogenies of contemporaneous species could be reconstructed using phenetic methods. Because morphological distance measurements are continuous data, analyses using numerical techniques is appropriate. Statistical Procedures General procedures All ofthe following statistical procedures were performed consistently across groups (Table 2.1) analyzed simultaneously (or as subsets), as the research questions demanded. However, specific statistical adjustments were required for the within- Palaemonetes comparison and the within-P. kadiakensis analyses (detailed below). In order to preserve allometric scaling effects among characters, alleviate effects of heteroscedasticity (Bookstein et al, 1985), and produce a scale-invariant covariance matrix, all character distances were converted to natural logarithms (Jolicoeur, 1963) prior to fiirther analysis. 19 Principal components analysis (PCA) was preformed on a covariance matrix of log-transformed data to determine possible structural relationships among the variables (Bookstein et al., 1985; Digby and Kempton, 1987). Individuals, or groups of individuals, were not assigned to any a priori categorical group, prior to running the PCA. And, the variation of all 30 characters was examined simultaneously for all taxa. Estimates ofthe relative allometric coefficients of individual characters, with reference to the general size-vector (PCI) were produced (Table 2.2). The populations were then subjected to discriminant function analysis (DFA). This procedure minimizes variation within designated groups and maximizes the variation between groups (Wiley, 1981). In order to discriminant groups on the basis of shape rather than size, size was regressed out ofthe data set prior to running a second DFA. Mahalanobis distances (D^) were calculated among groups, in the fijll multivariate space. Size was partitioned out ofthe data set by regressing each log-transformed character on PCI, which is typically the major size axis for biological forms. This provided the ability to do a size-free analysis (SF) and allowed the biological patterns in form to be interpreted as size-free shape contrasts. Centroids and confidence ellipses were generated by group to simplify visualization ofthe PCA plots, the DFA plots, and the central tendency ofthe distributions. The character correlations were plotted as vector diagrams (Wright, 1954; Strauss, 1985). Because DFA coefficients (weights) are not necessarily biologically interpretable it was necessary to do the vector analysis in this way. Phenetic clustering ofthe taxonomic groups was performed on Mahalanobis distances using a clustering algorithm (UPGMA) ofthe size-free DFA. This procedure allowed an interpretation ofthe distance relationships among all 20 characters in the full multivariate space, assuming equal rates of evolution for the characters. To evaluate taxonomic relationships, releasing the constraint of equal rates of evolution, additive dichotomous trees (neighbor-joining dendrograms) were constructed to explore Leander, Macrobrachium, Palaemon and Calathaemon as outgroups. All statistical analyses were performed on an IBM PC using commercial Matlab® (1992) routines and Matlab® routines (m files) generously provided by R. E. Strauss. The PCA, DFA, SF, vector plots, allometries, distance phenogram diagrams, and neighbor-joining trees were all produced using Matlab® m files. Influence of gender on character discrimination Sexual dimorphism in the full multivariate character space was assessed to determine whether males and females could be pooled, within taxa and populations, for fiirtheranalysis . However, due to small sample size of some groups (e.g., type specimens), sexual dimorphism within species and within taxa could not be statistically tested using a MANOVA. However, there is indirect inference that does not allow the rejection ofthe null hypothesis that males and females represent a common population in multivariate character space. The direct inference extends from that fact that: (1) a one-way MANOVA on a testable subset of organisms (i.e., P. kadaikensis) did not indicate any sexual dimorphism, (2) active researchers investigating the Palaemonidae have not uncovered any documented dimorphism in taxa studied (pers. comm., N. Strenth). It is a reasonable, yet an untestable inference that there is no significant sexual dimorphism among study organisms used in this work, but the question warrants further separate consideration. 21 Comparison of species within the genus Palaemonetes The genus Palaemonetes was examined as a single taxon, assuming the monophyly of relationships of North American species supported in recent literature. However, P. antromm was found in this study to be significanfly divergent from all others, and a separate series of analyses were run after removing P. antromm, in some cases. This was done in order to further define the internal relationships among the remaining taxa. Comparison within species of P kadiakensis Discrimination and phenetic relationships within species P. kadiakensis was not possible on the full data set because the number of characters exceeded the number of individuals. Character sorting was necessay to determine those characters which were most influential in character discrimination and such characters were retained for the multivariate analyses. Characters were evaluated using Rao's V, to determine which characters contributed most to the discrimination within the group (Tabachnick and Fidell, 1989). The Rao's V procedure resulted in a final character set reduction to 30 distance characters (and for all taxa, 2) which was entered into a DFA. Multivariate static allometric coefficients (Om) were scaled with respect to PCI by estimating the loadings (coefficients) on PCI for each character and scaling the loadings to a mean of 1.0 (Houck et al., 1990; Houck 1992; Hutcheson et al. 1995; Houck and OConnor, 1998). Evaluation of character contribution to discrimination The Rao's V procedure was applied to: all taxa, all species oi Palaemonetes (excluding P. antrorum), and to P. kadiakensis only. This was done to evaluate the relative importance of characters at various taxonomic levels. 22 CHAPTER m RESULTS AND DISCUSSION Influence of Gender on Character Discrimination An initial principal component analysis (PCA) of characters for males separate from females was completed on all taxa studied in order to determine whether there was character variation that could be attributable to sexual dimorphism. The convex hulls ofthe PCA, by gender, exhibited a significant amount of overlap, with the centroids relatively close together (Fig. 3.1). The centroids on PCI differed by approximately one standard deviation (SD) unit in the full space, and on PC2 by approximately 0.25 SD units. This indicates that there is little or no evidence for significant sexual dimorphism in this sample of organisms and thus data frommale s and females was pooled in subsequent analyses. A MANOVA was attempted, to establish an F statistic for gender differences among all taxa in the fiill space, but the individual sample sizes ofthe taxa were not large enough to contribute to a meaningful comparison. Analysis of All Taxa Examining all taxa studied, the first principal component (PCI) accounted for 67.5% ofthe total variation in the data set. PCI is a significant size vector as indicated by the magnitude and direction ofthe loadings and an apparent size trajectory that is correlated with the first eigenvector (Fig. 3.2). The plot of PC2 versus PCI suggested that Leander is the apparent outgroup ofthe taxa studied, with the generd Macrobrachium, Palaemon and Calathaemon positioned well within a cluster ofthe remaining taxa (Fig. 3.2). Separation of groups in multivariate PCA space is more clearly visualized by the centriod plot (Fig. 3.3). The centroid analysis indicated thai Macrobrachium is the largest ofthe taxa examined and P. antromm is this smallest. This is 23 All Taxa ^,''' \ Males \^ ; \ • Females fsv O CJ cu 1 ^ ^" ^ - / '' PCI (67.5%) Figure 3.1. Convex hulls and centroids indicating the discrimination among males and females, for all taxa studied. PCI, the major size axis, accounts for 67.5% ofthe variation in all taxa examined, and PC2, a major shape axis, accounts for 10.3%. PCI ranges from 0.0-6.0 standard deviation units, and PC2 ranges from 0.5-1.5 standard deviation units. 24 All Taxa Calathaemon / Freshwater Macrobrachium u -'Palaemon Marine Ijeander PCI (67.5%) Figure 3.2. Relationship between the first two axes of discrimination as determined by principal component analysis; convex hulls and centroids for all taxa studied The marine grouping consists of Leander, Macrobrachium, Palaemon, Palaemonetes hiltoni, P. intermedius, P. pugio, P. vulgaris plus P. paludosus and P. suttkusi which are freshwater inhabitants. The freshwater taxa consist of Calathaemon, P. antromm, P. cummingi, P. hobbsi, P. kadiakensis (six populations), P. lindsayi, and P. texanus. The dotted line indicates a separation ofthe basically fresh and salt water groups. The bold convex hulls represent genera that are possible outgroups. PCI, the major size axis, accounts for 67.5% ofthe variation in all taxa examined and PC2, a major shape axis, accounts for 10.3%. PCI ranges from 0.0-6.0 standard deviation units, and PC2 ranges from - 0.5-1.5 standard deviation units. 25 All Taxa Calathaemon Freshwater . P. antrom. Marine u ' ' p. paludosus sunkbt^- O ^ J*alaemo\ u .-'P.hiitom ' P^^^ Macrobrachium OH P. intermedins P. vulgaris Leander PCI (67.5%) Figure 3.3. Centroids and one standard deviation about the centroids as determined by principal component analysis for all taxa studied. Major freshwater species oi Palaemonetes from North America mainly fall into a cluster above the line of separation (—) between fresh and salt water species. No salt water species (including brackish water species) fall within the freshwater cluster. The bold ellipses represent Macrobrachium, one ofthe possible outgroups, and P. antrorum an aberrant subterranean species. PCI, the major size axis, accounts for 67.5% ofthe variation examined, and PC2, a major shape axis, accounts for 10.3%. PCI ranges from 0.0-6.0 standard deviation units, and PC2 ranges from - 0.5-1.5 standard deviation units. 26 corroborated by the univariate measurements of distance characters (Appendix). This analysis also suggests a separation ofthe marine from the freshwater species, v.ith the exception of P. paludosus and P. suttkusi which are intermediate between the marine species Macrobrachium and the marine species P. hiltoni and P. intermedius. An important fmding is that in no circumstance were any marine or brackish water taxa grouped among the freshwater groups. A possible interpretation is that the origin of this subset (freshwater Palaemonetes) ofthe Palaemonetidae had its origin in salt water and that P. paludosus and P. suttkusi retained some ofthe marine ancestral characters. The problem with this interpretation is that size pla\ s such a significant role in the separation of these taxa, and separation on size alone could account for this pattern. The univariate measurements of characters (Appendix) indicate that P. paludosus and P. suttkusi are relatively large members of this group. PC2 accounted for 10.3% of residual variation and PC3 accounted for 6.5%. PC2 and PC3 were major shape axes with potentially some unpartitioned residual size variation. A plot of PC3 versus PC2 (Fig. 3.4) indicated that Leander was separated from all other taxa on the major shape axes, and that P. antromm was discriminated from all other congeners oi Palaemonetes. The centroid analyis of PC3 versus PC2 offered no additional interpretive value (Fig. 3.5). Vector correlations ofthe first two major PC axis of discrimination (Fig. 3.6) indicated that no single character discriminated among the groups but that the characters form a uniform cluster which is heavily size loaded. Vector correlations ofthe major shape axes (PC2 and PC3) indicated that a rosette of characters contribute to the shape discrimination oi Leander and other taxa (Fig. 3.7). When examining the first two axes of discrimination of categorical groups, using discriminant fimction analysis (DFA), three initial clusters arose (Fig. 3.8): (1) P. antromm and Calathaemon; (2) Leander; (3) all other taxa. Interestingly, P. antromm and Calathaemon were collected from the same geological formation; an 27 All Taxa Calathaemon p. antromm Leander m Cu P. intermedius PC2(10.3%) Figure 3.4 Convex hulls and centroids indicating discrimination due to the two major shape axes as determined by principal component analysis for all taxa studied. PC2 accounted for 10.3% ofthe variation due to shape, while PC3 accounted for 6.5%. PC2 is scaled from -0.5-1.5 standard deviation units, and PC3 is scaled from -1.0-1.5 standard deviantion units . 28 All Taxa p. antrorum o Calathaemon Leander tn CO m cu P. intermedius PC2(10.3%) Figure 3.5. Centroids and one standard deviation about the centroids, due to the two major shape axes, as determined by principal component analysis, for all taxa studied. PC2 accounted for 10.3% ofthe variation due to shape, while PC3 accounted for 6.5%. PC2 is scaled from -0.5-1.5 standard deviation units, and PC3 is scaled from -1.0-1.5 standard deviation units. 29 cu c o CJ t O Correlation with PCI Figure 3.6. Vector plot showing character correlations. Arrows indicate the directionality of characters. In this plot all characters contribute relatively equal amounts to the character separation (see Figs. 3.2, 3.3). The length ofthe arrows represent the magnitude of contribution to group separation. PCI and PC2 axes range from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 30 m CJ 0. O CJ Correlation with PC2 Figure 3.7. Vector plot showing character correlations ofthe major shape axes. A rosette of characters contribute to the shape discrimination of all taxa (see Figs. 3.4,3.5). Both the PC2 and PC3 axes range from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 31 All Taxa / 'Sj\J\*// X • 1F7 wf*^A * -^ o^ Q P. antrorum ^ ^ Calathaemon DFl (37.5%) Figure 3.8. Plot ofthe major shape axis (DF2) versus the major size axis (DFl), as determined by discriminant fimction analysis for all taxa studied. DFl (size) accounts for 37.5% and DF2 (shape) accounts for 24.4% ofthe variation amoung taxa. The range of DFl is -2.0-2.0 standard deviation units, and DF2 ranges from -2.5-1.5 standard deviation units. 32 artesian well on the campus of Southwest Texas State University in San Marcos, Hayes Co., Texas and from Ezelle's Cave, San Marcos, respectively. Examination ofthe DFA centroid plot offers additional discrimination (Fig. 3.9). The marine and brackish water species {Macrobrachium, Palaemon, P. hiltoni, P. intermedius, P. pugio, P. vulgaris) and the freshwaterspecie s P. paludosus form one group while the majority of freshwater species cluster together. A plot of DF3 versus DF2 fiirther discriminates Calathaemon from P. antromm (Fig. 3.10). Centroid plots for the same relationship gave no further resolution and are not shown here. Vector analysis resulted in rosette formations for DF2 versus DFl (Fig. 3.11) and DF3 versus DF 2 (Fig 3.12) with no single character discriminating among groups. The character evaluation using Rao's V, indicated that ofthe 30 final characters used in the analysis the maximum number of discriminatory characters was not additive, but that the maximum discrimination power per character occurs at approximately five (Fig 3.13). As the number of characters are added to the data set, the rate of accumulated information content does not increase (Fig. 3.14). The characters which were discriminatory in the analysis (for all taxa evaluated simultaneously) are given in Table 3.1. The size-free discriminant analysis ofthe first two major axes of discrimination (Fig. 3.15) and the centroid analyses (Fig. 3.16) were qualitatively similar in information content to the analyses using size-uncorrected discriminant function analysis (Fig. 3.9). The conclusions reached were the same: the marine and brackish water species {Macrobrachium, Palaemon, P. hiltoni, P. intermedius, P. pugio, P. vulgaris) and the freshwaterspecie s P. paludosus form one group, while the majority of freshwaterspecie s cluster together. The subterranean species P. antromm and Calathaemon group together and exhibit certain qualitative characteristics similar to the marine species (i.e., antermal morphology; marine species have the unfused portion ofthe antennae longer than the fiised segments). 33 All Taxa '. intenvedius P paludosui ^—i^acrobrachium w Freshwater /' Phiitom (£) P- vulgaris CM I bu I Q I I / Marine p. antrorum ! O Leander • 4 Calathaemon DFl (37.5%) Figure 3.9. Centroids, and one standard deviation about the centroids, as determined by discriminant function analysis for all taxa studied. DFl (size) accounts for 37.5% and DF2 (shape) accounts for 24.4% ofthe variation amoung taxa. The range of DFl is -2.0- 2.0 standard deviation units, and DF2 ranges from -2.5-1.5 standard deviation units. 34 All Ta Freshwater Leander \ P. antrorum od Q Q Subterranean N \ ^ Calathaemon Marin• e \ \ DF2 (24.4%) Figure 3.10. Plot ofthe major shape axes, as determined by discriminant function analysis for all taxa studied. DF2 (primary shape axis) accounts for 24.4% and DF3 (secondary shape Jixis) accounts for 8.6% ofthe variation amoung taxa The range of DF2 is -2.0-1.0 standard deviation units, and DF3 ranges from -2.0-2.0 standard deviation units. 35 All Taxa 24 30 obu 17 G O ^11 J2 V Um Um 20^^ o U ^ 12 Correlation with DFl Figure 3.11. Vector plot showing character correlations of DF2 (shape axis) and DFl (size axis). Plot indicates that a rosette of characters contribute to the shape discrimination of all taxa (Figs. 3.8, 3.9). Numbers represent the major characters that discriminated taxa. DFl ranges from -1.0-1.0 standard deviation units and DF2 ranges from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 36 O oc 1 k. O Correlation with DF2 Figure 3.12. Vector plot showing character correlations contributing to character variation resulting in shape discrimination as determined by discriminant function analysis, for all taxa studied. Arrows show major characters contributing to the separation of taxa (Fig. 3.10). Both axes represent shape components, with DF2 ranging from - 1.0-1.0 standard deviation units, and DF3 ranging from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 37 /A. All Taxa 4.5 JJ X) 4.0 .2 ed > k. 3.5 a. Vi O 03 3.0 CJ > E 2.5 3 2.0 1 1 1 1 1 10 15 20 25 Number of variables included Figure 3.13. Plot ofthe cumulative Rao's V values for all taxa studied. The maximum discrimination power per character occurs with the fifdi character. 38 70 All Taxa 60 Vi o 50 OC > 40 3 30 E 3 CJ 20 10 0 10 15 20 25 Number of variables included. Figure 3.14. Cumulative Rao's V values for all taxa studied. The rate of accumulated information content per character decreases with the addition of characters. 39 Table 3.1. Rao's V hierarchy for the 26 characters used in the analvses of all taxa. 1. Dactyl Leg 11 2 Corpus Leg 11 3. Telson M x M 4. Corpus Leg 1 5. 3rd Segment Antenna Peduncle 6. Merus Leg IV 7. Telson L x L 8. 2nd Segment Antenna Peduncle 9. Corpus Leg V 10. Ischium Leg IV 11. 3rd Maxili Propodus/Dactyl 12. Carapace Length 13. Corpus Leg II 14. Dactyl Leg I 15. Propodus Leg V 16. Dactyl Leg IV 17. Ischium Leg I 18. 3rd Maxili Carpus 19. Telson, Base to middle spines 20. Telson, Middle spines to lower spines 21. Ischium Leg V 22. Propodus Leg IV 23. Merus Leg V 24. Merus Leg 1 25. Propodus Leg I 26. Ischium Leg II 40 All Taxa f — -Calathaemon/ . Marine ' 4 Leander P. antrorum Subterranean t J 00 Freshwater SFl (38.0%) Figure 3.15. Plot ofthe secondary axis of discrimination (SF2) and the primary axis of discrimination (SFl), as determined by size- free discriminant fimction analysis for all taxa studied. SFl accounts for 38.0% ofthe variation, and SF2 accounts for 24.8%. SFl ranges from -2.0-2.0 standard deviation units, and SF2 ranges from -1.5-2.5 standard deviation units. 41 Figure 3.16.Centroidsandonstandardeviationaboutthcentroid SF2 (24.8%) taxa studied.SFlaccountsfor38.0%ofthevariation,and2 units, andSF2rangesfrom-1.5-2.5standardeviation. accounts for24.8%.SFlrangefrom-2.0-2.0standarddeviation as determinedbysize-freediscriminantfunctionanalysiforall All Taxa (Z) Leander1 ^ Marine/ CD # / / / / \^^/ SFl (38.0%) 42 1 1 / / Subterranean Freshwater p. antrorum Calathaemon r A size-free cluster phenogram created by using unweighted pair group method using arithmetic averages (UPGMA) ofthe D^ unrooted (Fig. 3.17), suggested three hypotheses. (1) The marine taxa (plus P. paludosus, freshwater) form a cluster. (2) All ofthe remaining freshwater taxa cluster together, (3) A size- free cluster phenogram created using unweighted pair group method using arithmetic averages the freshwater cluster includes the subterranean species {Calathaemon, P. antromm and P. cummingi). A neighbor-joining unrooted cluster phenogram (Fig. 3.18) resulted in a grouping that included all ofthe freshwaterspecies , with the exception of P. paludosus. P. paludosus clustered with populations of marine and brackish water origin. The subterranean populations {Calathaemon, P. antromm and P. cummingi) form a subcluster among the freshwatertaxa , indicating a possible convergence in morphological form due to shared habitat constraints or the release of them. This suggests the hypothesis that the marine taxa are ancestral to the freshwater taxa and the subterranean taxa are highly derived freshwater forms. P. paludosus (freshwater in habitat) groups with the marine forms indicating that it may represent a possible second invasion of freshwater,bu t retaining many ofthe ancestral marine characteristics. If this is correct, the genus Palaemonetes is paraphyletic. A neighbor-joining phenogram, rooted at the logical outgroup as determined by the PCA analysis (i.e., Leander) (Fig. 3.19) resulted in the positioning of Palaemon and Macrobrachium well within the Palaemonetes phenogram and between the freshwater and marine clusters. These two taxa are considered to be brackish water inhabitants. These findings may indicate that Palaemon and Macrobrachium may be large marine relatives that are closely allied to Palaemonetes and are convergent to one another due to the shared habitat or shared ancestry. This is consistent with the hypothesis of Pereira (1989) that Palaemonetes is a sister taxon to Macrobrachium. 43 I *«» •^ it _N 'v: o c o u U o rj mm SB a U rf ee U ce « o a 6 8 a S o o oo *p •2 12 rr a er r o. > =-^ ra c rr a u ee o Q w .a o u J-) 3 Xi X CQ > S IJ ^ s: :3 1 to -S u •2 -2 1 arin e 5 5 •2* -2 e>o 2 o !^ 2 ^ SS •^ 1 s: a ^ o 2 'i ^1 ^ .2 -O c: <3 P "^ "^ .o S: -^ «M ^•§ 1 1 1 o a eo i«c •3 c: -:«: ^ -i: S: V) 5 OD c^ cC a^ =i; t^C^ C^'a.;G.^C.^ C^i^Ci^a,^ ^ c, JSu2 l-l O O u t/i a B 5to O g t/i c« iB O o tZ) O c c« BJ u o a> CO op o ^ o 44 c O 03 CQ ^ CJ B es .Si 3 C/3 Vi C3 ed oo c o r- CO c« IS Vi a: C a << •% a CJ o a c^ cd u OO a •^ o •^ c 1 - CJ e x; o ^ Q. Cu o CJ 00 «>c ^ i o 5 -2 X .2 ^ G _; «^* O o > a ^- 2 u > O a" = is X) .2 j:; ^ "3 «<0 o .2P *S <3 I S3 c a.; "^ 1 5 ~ -c -a I L_L 1 1 -«—> . . •Ac. . -^ ^ .2 o 5., =^ Ci, ft, ^ I I I •^' > p; c 5 ~ rS Q 2 1 j-a Si — .:r '^ a.; ^ c D V od l-r 00 45 c es CJ Ci CJ s es :: es Vi c u Cd ^ 'X ^ Q 73 <3 CJ o o T3 U ''B Vi cd X Cd cd O 00 U c a CJ o o JB B Cd { o o es Vi U o a, ao o U IS ee bo J H •H u Vi ee G «<5 a «> S >> CJ .2 u, 1 «^ CJ 2 -^ -J>c a (4-1 I 5 S3 <<5 o 1 1 p I -AC a I L •Ac I ^ a, (^ a Q 00 J_ I o G ««3 CJ x: <3 So 00 Si. Si, c ;s «>5 0. u- 3 ^ <5 o p -J :c op 00 3 -^ o ON rn 2i 3 00 fe 46 Comparison of Species within the Genus Palaemonetes A PCA was preformed on the genus Palaemonetes, by gender (Fig. 3.20) in order to determine whether there was within-genus character variation that could be attributed to sexual dimorphism. The convex hulls ofthe PCA, by gender, exhibited a significant amount of overlap, with the centroids relatively close together. The centroids on PCI differed by approximately one standard deviation (SD) unit in the full space, and on PC2 by no measurable difference. This indicates that there is little evidence for sexual dimorphism in the size component and no sexual dimorphism on the shape component of this sample of organisms. Since the centroids and convex hulls of this plot did not indicate any sexual dimorphism in character space, the genders were pooled for further analysis. A MANOVA was attempted, to establish an F statistic for gender differences among Palaemonetes, but the individual sample sizes ofthe taxa were not large enough to contribute to a meaningful comparison. The PCA plot indicated that PCI accounted for 70.8% ofthe variation in the characters (Fig. 3.21). The loadings of PCI were all positive, indicating that PCI is a general size vector. Species oi Palaemonetes fromNort h America can be separated into three groups based on distance measurements. The first discrimination is represented by P. antromm, the aberrant subterranean species which Holthuis (1952) placed in the subgenus Alaocaris. The freshwater species {P. cummingi, P. hobbsi, P. kadiakensis, P. lindsayi, P. mexicanus, P. suttkusi, and P. texanus) form the third group (group A). A marine group (group B) consists ofthe species, {P. vulgaris, P. intermedius, P. pugio, and P. hiltoni) and P. paludosus. The centroid analysis further clarified the central tendencies ofthe groups (Fig. 3.22). PC2 accounted for 8.4% ofthe character variation species. Shape appeared to be the largest contributing factor to separation within the freshwater and marine groups. A plot of PC3 on PC2 (Fig. 3.23) maintained the three groups but failed to contribute additional information of separation within the groups. A character 47 00 PCI (70.8%) Figure 3.20. Convex hulls and centroids indicating the discrimination among males and females, for species of Palaemonetes from North America. PCI, the major size axis, accounts for 70.8% ofthe variation in all taxa examined, and PC2, a major shape axis, accoimts for 8.4%. PCI ranges from 0.0-6.0 standard deviation units, and PC2 ranges from - 0.4-1.6 standard deviation units. 48 Palaemonetes p. antromm ^ o^ 00 Freshwater ON PCI (70.8%) Figure 3.21. Relationship between the firsttw o axes of discrimination as determined by principal component analysis; convex hulls with centroids for populations oi Palaemonetes from North America. The freshwater group (Group A) consists of Calathaemon, P. antromm, P. cummingi, P. hobbsi, P. kadiakensis (six populations), P. lindsayi, and P. texanus. The marine group (Group B) consists oi Leander, Macrobrachium, Palaemon, Palaemonetes hiltoni, P. intermedius, P. pugio, P. vulgaris plus P. paludosus and P. suttkusi which are freshwater inhabitants. The dotted line indicates a separation ofthe basically fresh and salt water taxa. PCI, the major size axis, accounts for 70.8% ofthe variation and PC2, a major shape axis, accounts for 8.4%. PCI ranges from 0.0-6.0 standard deviation units, and PC2 ranges from - 0.5-1.5 standard deviation units. 49 Palaemonetes p. antromm P.cummingi vulgaris P. pugio P. intermedius Marine PCI (70.8%) Figure 3.22. Centroids and one standard deviation about the centroids as determined by principal component analysis for species oi Palaemonetes from North America PCI, the major size axis, accounts for 70.8% ofthe variation, and PC2, a major shape axis, accounts for 8.4%. PCI ranges from 0.0- 6.0 standard deviation units, and PC2 ranges from - 0.5-1.5 standard deviation units. 50 Palaemonetes. P. antromm m Marine Freshwater PC2 (8.4%) Figure 3.23. Convex hulls and centroids indicating the discrimination among species oi Palaemonetes from North America, due to the two major shape axes. PC2, the major shape axis, accounts for 8.4% ofthe variation examined, and PC2, a major shape axis, accounts for 6.3%. PC2 ranges from -0.5-1.5 standard deviation units, and PC3 ranges from - 0.8- 0.6 standard deviation units. 51 correlations vector plot suggests that all the characters contributed relatively equally to the separation ofthe groups (Fig. 3.24). Discrimmant function analysis (Fig. 3.25) suggests support for the division of the populations oi Palaemonetes into two groups: (1) a marine cluster (plus P. paludosus); (2) the freshwater groups, and P. antrorum as an outlier. Complete discrimination is expressed on both the major size axis and the major shape axis. Centroid analysis provides no more information concerning the groups and is not shown here. DFl accounts for 41.8% ofthe variation while DF2 is responsible for 24.3%. The vector plot of DF2 versus DFl (Fig. 3.26) suggests that the major characters of discrimination form a rosette by which the groups separate. A second discriminant function analysis, DF3 versus DF2 (Fig. 3.27), was completed to determine any further separation within the groups. The freshwater species remain in a tight cluster with no discrimination within the group. However, within the marine group the convex hulls are more resolved. Palaemonetes vulgaris clearly separates from the other marine species. Palaemonetes hiltoni also separates from other members ofthe group. The remaining species only marginally separate. In this DFA, DF3 accounts for 9.9% ofthe variation. An interesting issue is that P. antromm clusters with the marine group. Based on the position of P. antrorum and P. paludosus in the neighbor-joining phenogram (Fig. 3.19), it suggests that these two freshwater species have retained similar marine ancestral shape characters. The importance of this is, that it offers additional inference that the freshwater Palaemonetes is not monophyletic. A plot ofthe centroids and one standard deviation (Fig. 3.28), on DF2 and DF3 axes, suggests that the individuals within the marine group do indeed separate with no overlap of characters at one standard deviation. The vector plot indicates the characters that separate the different species oi Palaemonetes in the second DFA (Fig. 3.29). 52 Palaemonetes y^^ 1 oCN Q. ^rrr ^^ vit h tion v (0 0) v^^12 oo 11 Correlation with PCI Figure 3.24. Vector plot showing character correlations contributing to character separation among species of Palaemonetes from North America (Figs. 3.21, 3.22). Arrows show a directional component with all characters contributing relatively equal amounts of variation to the separation of taxa. Both the PCI and PC2 axes range from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 53 Palaemonetes y m Ci Freshwater P. paludosu^ O P. antromm / Marine DFl (41.8%) Figure 3.25. Plot ofthe major shape axis (DF2) verus the major size axis (DFl), as determined by discriminant function analysis for all species oi Palaemonetes from North America. DFl accounts for 41.8% of the variation within the North American Palaemonetes, while DF2 accounts for 24.3%. DFl ranges from -3.0-2.0 standard deviation units, and DF2 ranges from -2.0-1.5 standard deviation units. 54 Correlation with DF 1 Figure 3.26. Vector plot showing character correlations among species oi Palaemonetes from North America. DF2 (shape axis) and DFl (size axis) indicates that a rosette of characters contribute to the shape discrimination of all taxa (Fig. 3.25). Numbers represent the major characters that are used in the discrimination of taxa. DFl ranges from -1.0-1.0 standard deviation units, and DF2 ranges from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 55 Palaemonetes p. pugio P. paludosus P. antrom< P. intermedins cn Marine Freshwater p. vulgaris DF2 (24.3%) Figure 3.27. Plot ofthe major shape axes, as determined by discriminat function analysis for populations oi Palaemonetes from North America. DF2 accounts for 24.3% ofthe variation and DF3 accounts for 9.9%. DF2 ranges from -2.0-1.5 standard deviation units, and DF3 ranges from -3.0- 2.0 standard deviation units. 56 Palaemonetes r-^-^v^. pugio p. paludosus\^_j P. intermedins bu Q Freshwater r P. vulgarise / /•' DF2 (24.3%) Figure 3.28. Centroids and one standard deviation about the centroids for the major shape axes, as determined by discriminant analysis, for species oi Palaemonetes fromNort h America. DF2 accounts for 24.3% ofthe variation and DF3 accounts for 9.9%. DF2 ranges from -2.0-1.5 standard deviation units, and DF3 ranges from -3.0-2.0 standard deviation units. 57 Palaemonetes 29 Q 5 28 ^y^ 30 c o at i ^^^^>> 10 rre l o ''^'^^*r^ 8 Correlation with DF2 Figure 3.29. Vector plot showing character correlations of DF3 (minor shape axis) versus DF2 (major shape axis) among species oi Palaemonetes from North America (Figs. 3.27, 3.28). The plot indicates that a rosette of characters contribute to the shape discrimination studied. Numbers represent the major characters that are used in the discrimination of taxa. DFl ranges from -1.0-1.0 standard deviation units and DF2 ranges from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 58 The character evaluation using Rao's V suggests that, ofthe 30 final characters used in the analysis, the maximum number of discriminatory characters was not additive and the maximum discrimination power per character occurs with the addifion ofthe fifth character (Fig 3.30). As the numbers of characters were added to the data set, the rate of accumulated information content did not increase (Fig. 3.31). The characters that were discriminatory in the analysis (for all taxa evaluated simultaneously) are given in Table 3.2. Size was removed from the data set and a size-free DFA was completed. The size-free DFA suggests a consistent separation ofthe same groups as above. However, the SFl vector accounts for only 42.3% ofthe variation (Fig. 3.32). The second and third axes ofthe size-free DFA, do not suggest further separation within either group (Fig. 3.33). A size-free discriminant fimction analysis cluster phenogram based on the D was created using unweighted pair group method using arithmetic averages (UPGMA). This unrooted phenogram (Fig. 3.34) suggests marine taxa (plus P. paludosus, freshwater) cluster together. The remaining freshwater taxa cluster together including the subterranean species {Calathaemon, P. antromm and P. cummingi). A result of a neighbor-joining unrooted cluster phenogram (Fig. 3.35) is the grouping of all freshwaterspecie s except P. paludosus. P. paludosus clusters with populations of marine and brackish water origin. The subterranean populations {P. antromm and P. cummingi) form a subcluster among the freshwatertaxa . The marine taxa clustered together. Comparison of Discrete Characters As stated in the introduction, several discrete characters have been used in an attempt to resolve the systematic relationships within the North American species of the genus Palaemonetes. The egg size and number are consistent in all species examined, that is the freshwater forms have larger but fewer eggs than are found in 59 Palaemonetes 4.0 JJ .0 .5 I3.5 CJ Q. > Vi §3.0 OC CJ > Cd 1 3 2.5 S3 u 2.0 1 1 1 I « 0 10 15 20 25 Number of variables included Figure 3.30. Plot ofthe cumulative Rao's V values for all species of Palaemonetes from North America showing the maximum discrimination power per character occurs at approximately the fifth character measured. 60 Vi o Cd > 3 E 3 0 10 15 20 Number of variables included Figure 3.31. Cumulative Rao's V values for all species of Palaemonetes from North America showing that the rate of accumulated information content per character decreases with the addition of characters after the addition ofthe fifth character. 61 Table 3.2. Rao's V hierarchy the 30 characters used m the analyses oi Palaemonetes. Telson L x T. 2. Propodus Leg II 3. Telson M x M Corpus Leg II Corpus Leg V 6. Telson B X M 7. Corpus Leg II 8. Telson L x L Merus Leg IV 10. Propodus Leg IV 11, Ischium Leg V 12. Rostrum length 13. Carapace Length 14. Corpus Leg I 15. 3rd Maxili Propodus/Dactyl 16. Merus Leg V 17. Dactyl Leg 18. Ischium Leg IV 19. Propodus Leg I 20. Dactyl Leg IV 21. Ischium Leg II 22. 3rd Maxili Carpus 23. Ischium Leg I 24. Telson M x L 25. 3rd Segment Antenna Peduncle 26. 2nd Segment Antenna Peduncle 27. Corpus Leg IV 28. Dactyl Leg 29. Merus Leg I 30. Propodus Leg V 62 Palaemonetes Freshwater 00 Um p. antromm Marine SFl (42.3%) Figure 3.32. Plot ofthe second size-free discriminatory axis (SF2) vs. the first size-free discriminatory axis (SFl) for the North American Palaemonetes. SFl accounts for 42.3% of the variation, and SF2 accounts for 24.8%. SFl ranges from -2.0-3.0 standard deviation units and SF2 ranges from -2.0- 1.5 standard deviation units. 63 Palaemonetes P. pugio I p. paludosu p. vulgaris SF2 (24.8%) Figure 3.33. Plot ofthe third size-free discriminatory axis (SF3) vs. the second size-free discriminatory axis (SFl) for the North American Palaemonetes. SF2 accounts for 24.8% of the variation, and SF3 accounts for 10.2%. SF2 ranges from -3.0-2.0 standard deviation units, and SF3 ranges from -2.0 to 1.0 standard deviation units. 64 u CJ es CJ c es CJ es s u CJ o C/5 6 O o 'o O Ci O c o c 0) O o i5> o O c 51? '•D ^ b (D 3 o m 1- Co' co"" .«0 S2 CO yv^ Cd ^ u e c t/i es !._-,• o o o t/i IS o c J2 C3 o O 0) c« (N (U I N 00 O o 65 c es' Ci u' C/3 b.u1, o ao. SI OD C o Cd Vi 6C •^ C: C o ft. _^ S-i U a.; o Ci B ed s kd o . to U O 00 6 Vi o Ci d eei i u •o a G u a> .s ^ -c a, c Llau o C alg o Co . . Burlcs o . Va l Ve r , Menar d «n b> ^ 66 the marine species. While P. kadiakensis remains the exception, the freshwater species exhibit a reduced number of larval stages than the marine. The ftised segments ofthe antennular flagellum also serve as a discrete character to separate the marine and freshwater species. While P. antrorum and P. holthuisi are exceptions to this character, it may be they have retained ancestral characters that may benefit them in a tactile environment. The segmentation on the antennal scale in the form I zoea presents some questions. Palaemonetes kadiakensis, a freshwater species, exhibits a segmented antennal scale as do the marine species. Also, P. paludosus, another freshwater species, has an unsegmented antennul scale in the form I zoea. In all ofthe analyses performed in this study, P. paludosus always groups with the marine species even though the segmentation ofthe form I zoea is consistent with the freshwater species. The grouping ofthe freshwater species may be due to similarities in sediment load ofthe environments inhabited by this group of species. The movable exopod spination as a possible systematic character proposed by Strenth (1976) does have some support in that P. hobbsi and P. mexicanus group together and share the same movable expod condition. Palaemonetes lindsayi and P. suttkusi exhibit a lack of exopod spines and group together. However, the subterranean form P. antrorum, which shares the same exopod condition as do P. lindsayi and P. suttkusi. is grouped with the other subterranean species. The species P. texanus appears intermediate within populations oiP. kadiakensis; however, P. texanus exhibits the same condition ofthe movable exopod spine as does the P. hobbsi and P. mexicanus group. Larger sample sizes may help to resolve the relationships based on the condition ofthe movable exopod spine. Two additional characters were discussed in the introduction, rostrum dentition and appendix masculinae, as being considered as possible characters to resolve systematic questions of this group. As was pointed out, the variation within each species is large and overlaps among the different species. 67 Comparison within Species of P kadiakensis Six populations of P. kadiakensis fromwithi n the state of Texas were examined. In order to determine the effects of sexual dimorphism, a PCA was performed, by gender (Fig. 3.36). Examination ofthe convex hulls for PC2 versus PCI shows males and females overlap significantly suggesting a lack of sexual dimorphism within these populations of P. kadiakensis. The within-species comparison contained a larger sample than the comparison across all taxa or within the genus Palaemonetes. Because ofthe significant samples size ofthe pooled P. kadiakensis populations (N~60), a MANOVA was successfijlly applied to examine the degree of statistical difference between males and females in multivariate space. The F statistic indicate males do not significantly discriminate from females in the fiill character space (F 30,6= 0.43; p=0.94). Because this was found, data were pooled across gender for fiirtheranalysis . The PCA for the subset of P. kadiakensis was performed and PCI accounted for 78.1% ofthe variation (Fig. 3.37) within the characters. This suggests that most ofthe separation within these populations is due to size. Only 3.2% ofthe variation is accounted for by PC2. The centroid and one standard deviation plot for PCI versus PC2 (Fig. 3.38) more clearly defines the separation ofthe populations. With the exception of a minor overlap with the highly variable Tom Green Co. population, the Hidalgo Co. population is separate fromth e remaining five populations. The centroids of four ofthe six populations lay along the major size axis (PCI) and cluster together. It is interesting to note size plays an important role in the separation ofthe Hidalgo Co. and Val Verde Co. populations, which both are part ofthe Rio Grande River drainage system. The Val Verde Co. population exhibits marginal separation from the other populations with only a small amount of overlap between the ellipse for the Menard and the Val Verde Co. populations. There appears to be a large shape component v^thin the variation for 68 P. kadiakensis / \ ' \ / \ / \ , \ / \ / \ . ^ y—v / _ ©^ \ l^ •""•^ " \ rS C/ Males • ^Females C2 ( \ cu \ ' \ ^^^^ \ \ \^'''''^'^ \ L ^ \ \ •"" —. ___ \ \ •"* -^ 4# \ ""^^•---.^^^^ PCI (78.1%) Figure 3.36. Convex hulls and centroids indicating the discrimination among males and females, for populations of Palaemonetes kadiakensis. No sexual dimorphism was evident (FJQ^ = 0.43; p = 0.94). PCI, the major size axis, accounts for 78.1% ofthe variation in all taxa examined and PC2, a major shape axis, accounts for 3.2%. PCI ranges from 1.0-6.0 standard deviation units, and PC2 ranges from - 0.6-0.8 standard deviation units. 69 P, kadiakensis 11 Hidalgo Co. Val Verde Co. m" ^ ^^^ Menard Co. / • (^^ > 0^ \ .^ ' ^ • \Tom Green Co. \ .^^ w — ~ A^r . Burleson Co. Llano Co. ^"^^^^^^^ \ PCI (78.1%) Figure 3.37. Relationship between the first two axes of discrimination as determined by principal component analysis; convex hulls with centroids for populations oi Palaemonetes kadiakensis studied. PCI, the major size axis, accounts for 78.1% ofthe variation in all taxa examined and PC2, a major shape axis, accounts for 3.2%. PCI ranges from -1.0-6 standard deviation units, and PC2 ranges from -0.6-0.6 standard deviation units. 70 P. kadiakensis Hidalgo Co. Val Verde Co. ©^ rn Menard Co. cu Burleson Co Tom Green Co. PCI (78.1%) Figure 3.38. Centroids and one standard deviation about the centroids as determined by principal component analysis for populations of Palaemonetes kadiakensis studied. PCI, the major size axis, accounts for 78.1% of the variation in all taxa examined and PC2, a major shape axis, accounts for 3.2%. PCI ranges from 1.0-6.0 standard deviation units, and PC2 ranges from -0.6-0.6 standard deviation units. 71 the Tom Green Co. population. The vector plot for PC2 versus PCI (Fig. 3.39) suggests that all characters contribute relatively equal in the separation of these six populations. There is very little shape separation in PC3 versus PC2 (Fig. 3.40), each axis accounts for only about 3% ofthe variation. However, there is marginal separation ofthe Burieson Co. population. All other populations cluster with little separation of the centroids ofthe convex hulls. The centroid and one standard deviation plot for PC3 versus PC2 (Fig. 3.41) suggest the Burieson Co. population does separate from the remaining populations with no overlap. The vector plot (Fig. 3.42) suggests the most significant character in the separation ofthe Burleson Co. population is the ischium of leg one. DFl accounts for 86.4% ofthe variation, while DF2 accounts for 10.3%. All populations discriminate in DF2 versus DFl (Fig. 3.43). The DF3 versus DF2 (Fig. 3.44) plot also discriminate the six populations. However, the relationship ofthe populations from Hidalgo Co. and Menard Co. is seen to be farther apart, and the Burleson Co. population becomes more intermediate between the other two populations. However, this may be artifact of reduced sample size. The centroids and one standard deviation plot for DF3 versus DF32 do not provide any additional informative information and are not shown here. The character correlation vector plot of DF 2 versus DFl (Fig. 3.45) suggests the characters used in the separation of populations is directional, with no one character having a significant effect on the discrimination of this group of Palaemonetes. The vector plot for DF3 versus DF2 (Fig. 3.46) also suggests no one character has a significant effect. The size-free discriminant function strongly differentiates the six populations. SFl accounts for 67.4% ofthe variation while SF2 is responsible for 25.2% (Fig. 3.47). There is no additional information contributed by the centroid and the one standard deviation plot, so this diagram is not figuredhere . The vector plot (Fig. 3.48) is represented as a rosette with no single character having a 72 Ci Um Um o Correlation with PCI Figure 3.39. Vector plot showing character correlations. Arrows show a directional component with all characters contributing relatively equal amounts of variation to the separation of populations oi Palaemonetes kadiakensis (Figs. 3.37, 3.38). The PCI ranges from -1.0-1.0 standard deviation units, and PC2 ranges from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 73 P, kadiakensis Burleson Co. Hidalgo Co. OH Menard Co. Val Verde Co. Llano Co PC2 (3.2%) Figure 3.40. Convex hulls and centroids indicating the discrimination among all taxa due to the two major shape axes for Palaemonetes kadiakensis populations studied. PC2, the major size axis, accounts for 3.2% ofthe variation in all taxa examined and PC3, a major shape axis, accounts for 3.0%. PC2 ranges from -0.6-0.6 standard deviation units, and PC3 ranges from -1.0 to -0.2 standard deviation units. 74 P. kadiakensis Burleson Co Menard Co. cn cn U Tom Green Co Hidalgo Co. Val Verde Co. Llano Co. PC2(3.2%) Figure 3.41. Centroids and one standard deviation about the centroids due to the two major shape axes, as determined by principal component analysis, for populations oi Palaemonetes kadiakensis studied. PC2, the major shape axis, accounts for 3.2% ofthe variation in all taxa examined, and PC3, a minor shape axis, accounts for 3.0%. PC2 ranges from -0.6-0.6 standard deviation units, and PC3 ranges from -1.0 to -0.2 standard deviation units. 75 P, kadiakensis Ischium Leg I 29 um JZa. 13 \ J ^ 25 •| c o 2 ^^^% ""S -re l 16 ^^10 O U Correlation with PC2 Figure 3.42. Vector plot showing character correlations. Arrows show a directional component with all characters contributing relatively equal amounts of variation to the separation of populations oi Palaemonetes kadiakensis (Figs. 3.40, 3.41). PC2 ranges from -1.0-1.0 standard deviation units, and PC3 ranges from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 76 P. kadiakensis \ Val Verde Co. Llano Co. P Menard Co. cn ^ O Burleson Co. 0 CN 0 Hidalgo Co. Tom Green Co. DFl (86.4%) Figure 3.43. Plot ofthe major size axis vs. the major shape axis, as determined by discriminant function analysis for populations oi Palaemonetes kadiakensis studied. DFl accounts for 86.4% ofthe variation while DF2 accounts for 10.3%. DFl ranges from -1.5-1.5 standard deviation units, and DF2 ranges from -2.0-2.0 standard deviation units. 77 Hidalg oCo^ P. kadiakensis Val Verde Co. Burleson Co. ©^ Menard Co. CN m Q Tom Green Co. Llano Co. V DF2(10.3%) Figure 3.44. Plot ofthe major shape axes, as determined by discriminant function analysis for populations of Palaemonetes kadiakensis studied. DFl accounts for 10.3% of the variation, while DF2 accounts for 2.0%. DFl ranges from -2.0-2.0 standard deviation units, and DF2 ranges from -2.5- 1.5 standard deviation units. 78 Correlation with DFl Figure 3.45. Vector plot showing character correlations for populations of Palaemonetes kadiakensis. Arrows show a directional component with all characters contributing relatively equal amounts of variation to the separation among populations (Fig. 3.43). The DF2 ranges from -1.0-1.0 standard deviation units, and DF3 ranges from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 79 Correlation with DF2 Figure 3.46. Vector plot showing character correlations for populations of Palaemonetes kadiakensis. Arrows show a directional component with all characters contributing relatively equal amounts of variation to the separation among populations (Fig. 3.44). The DF2 ranges from -1.0- 1.0 standard deviation units, and DF3 ranges from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 80 P. kadiakensis vp- Tom Green Co. Burleson Co. Hidalgo Co. ©^ CN in CN CN \ 7 Menard Co. 17 ^ Llano Co. Val Verde Co. SFl (67.4%) Figure 3.47. Plot ofthe major discriminatory axis (SFl) and the secondary discriminatory axis (SF2), as determined by size-free discriminant function analysis for the genus Palaemonetes kadiakensis. SFl accounts for 67.4% ofthe variation, and SF2 accounts for 25.2%. SFl ranges from -2.0-2.0 standard deviation units, and SF2 ranges from -2.0-2.0 standard deviation units. 81 CN tu C CJ Um O U Correlation with SFl Figure 3.48. Vector plot showing character correlations. Arrows show a directional component with all characters contributing relatively equal amounts of variation to the separation among populations of Palaemonetes kadiakensis (Fig. 3. 47). The SF2 ranges from -1.0-1.0 standard deviation units, and SF3 ranges from -1.0-1.0 standard deviation units. See Table 2.2 for the list of characters by character number. 82 significant effect on the discrimination of these populations. The convex hulls for SF3 versus SF2 (Fig. 3.49) also discriminate all ofthe populations ofthe genus of Palaemonetes from North America. The Burleson Co. population moves into a more central position within the group of populations. And the Tom Green Co. and the Val Verde populations appear to shift the relationship between these two groups with Burleson Co. and the Menard Co. populations falling between the Tom Green and the Val Verde populations. The centroid and one standard deviation plot did not add any informative information. Rao's V (Fig. 3.50) was used to examine the value added by each additional character. Because ofthe close genetic relationship ofthe populations, more characters may be needed to see subtle within-species discrimination. The slope of this relationship begins an ascend greatly at 25 characters, and the information content at the maximum number of characters (i.e., 30) is likely due to accumulated noise and not accumulated character value. A major finding ofthe within-species comparison is the amount of discrimination evident among the six P. kadiakensis populations from Texas. The separation pattern follows the paradigm suggested for maximum evolution of closely related taxa. That is, if the populations are in restrictive drainage systems isolated for long periods ofthe populations have the potential to become independently dissociated from the marine ancestral forms in various and autonomous manners. The plot for the cumulative Rao's V (Fig 3.51) indicated that as characters were added to the Rao's V analysis the rate of information per character added decreased after the addition ofthe tenth character. The characters which were discriminatory in the analysis (for all taxa evaluated simultaneously) are given in Table 3.3. The unrooted UPGMA phenogram representing the size-free Mahalanobis distance analysis (Fig. 3.52) indicates a possible clustering ofthe Tom Green and the Val Verde populations and a second cluster containing the remaining populations. 83 p. kadiakensis Val Verde Co. Hidalgo Co. Burleson Co. ©^ m Ui Menard Co. Tom Green Co. V Llano Co. SF2 (25.2%) Figure 3.49. Centroids and convex hulls, as determined by size-free discriminant analysis for populations oi Palaemonetes kadiakensis studied. SF2 accounts for 25.2% ofthe variation, and SF3 accounts for 4.5%. SF2 ranges from -2.0-2.0 standard deviation units, and SF3 ranges from -2.5-1.5 standard deviation units. 84 0 10 15 20 25 Number of variables included Figure 3.50. Plot ofthe cumulative Rao's V values for all populaitons oi Palaemonetes kadiakensis studied showing the maximum discrimination power per character. See text for detailed discussion. 85 P. kadiakensis 60 ^^^ > 50 ^.^-^""^^ Vi ^^^'"""^ 1 40 y^ o - /^ Cumulativ e 20 - / 10 / 1 1 II 1 0 10 15 20 25 Number of variables included Figure 3.51. Cumulative Rao's V values for populations oi Palaemonetes kadiakensis studied showing that the rate of accumulated information content per character decreases with the addition of characters. 86 Table 3.3. Rao's V hierarchy the 30 characters used in the analyses oi Palaemonetes kadiakensis. 1. Corpus Leg V 2. Ischium Leg II 3. Corpus Leg 1 4. Ischium Leg 1 5. Telson B X M 6. Propodus Leg II 7. Dactyl Leg 1 8. Merus Leg V 9. 3rd Maxili Carpus 10. 3rd Maxili Propdus/Dactyl 11. Carapace Length 12. Rostrum length 13. Propodus Leg 1 14. 3rd Segment Antenna Peduncle 15. Corpus Leg II 16. Ischium Leg IV 17. Propodus Leg V 18. Merus Leg 1 19. Ischium Leg V 20. Dactyl Leg II 21. 2nd Segment Antenna Peduncle 22. Telson L x T 23. Telson M x M 24. Dactyl Leg IV 1 25. Corpus Leg II 26. Telson M x L 27. Telson L x L 28. Propodus Leg IV 29. Merus Leg IV 30. Corpus Leg IV 87 o , u o o a L> O o u a> o o a H "P u o o Z> U ce o •a a es I ce > u a 2 o u CoN o Cc« ^C/^5 * VB^ Q c« • ^^ o X) "do - o ^•^ JSa a o 2 o a> so a> u • N^"t C/5 o 0o0 88 A similar pattern of clusters is supported by the unrooted neighbor joining phenogram (Fig. 3.53). However, there appears to be a shift among the populations ofthe Burleson Co., Hidalgo Co., Llano Co. and Menard Co. cluster. 89 c o d •—: u u ^ a> o ^a^ Ic s oo 8 Rii r ip s amo r 1 . ounl y o f •S " a ^ = ^ H 4> O X) 1 1 '5 = *TO« >O a > L^ •—^ ,s •« 00 5- > *j ••» ^ S o o ? u G "^ o S ^3 a TO r** ZJ G2 -r^ 6 6 U o ^ ee a -a oinin g p laemone X 2 Z? Q o ^ JD= (4-O1 op Vi 5 G G ,0 V.CJ "S— "^ 3 2 2 c c:^ D cn u-^ rn aU>i 3 00 Pu 90 CHAPTER IV CONCLUSIONS Several consistent patterns emerge from this work: (1) Leander appears to be a valid outgroup for the North American species oi Palaemonetes. This possible ancestral connection along with the fact that there are extant marine species of this genus indicates that Palaemonetes likely arose in the marine environment, rather than arising in freshwater and re-invading the marine environment as hypothesized by Pereira (1989); (2) Calathaemon appears to be a highly derived species of Palaemonetes. In every analysis, Calathaemon appears within the genus Palaemonetes rather than separated from the group. This research provides several hypotheses for the systematic relationships of Palaemonetes from North America. Systematic reevaluation ofthe taxa indicate that there is little evidence to support the hypothesis by Bruce (1993) that P. holthuisi should be elevated to a new genus Calathaemon. It should remain in the genus Palaemonetes and the name of P. holthuisi restored to its original designation as stated in the rules of taxonomic nomenclature. The elevation of P. antrorum to the subgenus Alaocaris is not supported by this research and I feel P. antromm should remain within the genus Palaemonetes. Palaemonetes antromm appears to be a highly derived species within the genus Palaemonetes; however, it appears to be an intermediate sister taxon among other species in the genus. The group containing P. antrorum, P. cummingi, and Calathaemon require fiirther study to understand whether this group is indeed a monophyletic group. All three species live in a subterranean environment, so it is highly likely that this group exhibits convergent evolution rather than a close systematic relationship. The data in this study supports P. paludosus to be more closely related to the marine species than to the other freshwater groups, and fiirther suggests the freshwater species have arisen on at least two different occasions. However, P. paludosus exhibits the freshwater conditions associated with this group, egg size and 91 number, reduced larval stages, and the non-segmentation ofthe antennular scale in the form I zoea suggests P. paludosus is more closely associated with the other freshwater species oi Palaemonetes from North America. A possible explanation for the grouping of P. paludosus with the marine group may be parallelism from the retention of ancestral characters or convergent evolution due to similarities in sediment loads within their enviroimients. The data suggests the remaining freshwater species oi Palaemonetes form a monophyletic group representing a single invasion into freshwateran d this group has undergone radiation since that invasion. Therefore, fiirther study is needed to resolve the hypothesis by Sollaud (1923) and Strenth (1976) proposing North American freshwater Palaemonetes form a monophyletic group. Dobkin's (1965) proposal that P. kadiakensis is a recent immigrant into the freshwater environment is not supported. Palaemonetes kadiakensis appears to be ancestral to other freshwaterspecie s of this genus. The segmented antennule ofthe first zoea of P. kadiakensis would appear to be due to convergent evolution with the marine forms. The extended larval development stages exhibited by P. kadiakensis may in fact be a result of selection because ofthe wide spread and variable enviroimiental conditions found across the range of this species. The marine and freshwater taxa represent two divergent groups, probably separated at some ancient time and then radiating within their respective habitats. However, P. paludosus appears to be more closely related with the marine species than to other freshwatergroups . The neighbor-joining phenogram indicated that the marine species are probably ancestral to the freshwater species. This would suggest that the freshwaterspecie s have arisen twice and therefore do not form a cohesive group, as proposed by Sollaud (1923) and Strenth (1976). However, the remaining freshwater species oi Palaemonetes do appear to form a paraphyletic group however. Subterranean species fiirtherdiverge d in a convergent maimer to exploit that extreme habitat. 92 Several biogeographic patterns also emerged. The three subterranean species, P. antromm, P. cummingi, and the genus Calathaemon, are consistently grouped together. The two species from the eastern slope ofthe Sierra Madre Oriental Mountains in Mexico, P. mexicanus and P. hobbsi repeatedly appear to be sister taxa. While P. suttkusi and P. lindsqyi, both fromth e west ofthe Sierra Oriental Mountains form a natural group as well. The populations of P. kadiakensis appear to be broken into two groups depending on their geographic range. One group of P. kadiakensis, includes populations that are east and south of a line following the Balcones Escarpment in central Texas. The second group of populations of P. kadiakensis are found north and west ofthe line. Palaemonetes texanus, fromHaye s Co., Texas appears to separate the north and western populations from the eastern and southern populations. Within the out groups included in this study, Palaemon and Macrobrachium appear to be closely related; however, Leander appears to more closely aligned with the marine group of Palaemonetes than to either Palaemon or Macrobrachium. When Leander is included in the analysis, it always appears closely related to P. vulgaris. 93 LITERATURE CTFED Benedict, J. E. 1896. Preliminary descriptions of a new genus and three new species of crustaceans from an artesian well at San Marcos, Texas. Proceedings of the United States Natural History 18:615-617. Bookstein, F., B. Chemoff, R. Elder, J. Humphries, G. Smith, and R. Strauss. 1985. Morphometries in Evolutionary Biology. Special Publication 15, The Academy of Natural Sciences of Philadelphia. 277 pp. . 1989. Size and shape: A comment on semantics. Systematic Zoology 38:173-180. Broad, A. C. 1957. Larval development of Pa/aremowe/e5/7i/g/o Holthuis. The Biological Bulletin 112:144-161. Broad, A. C. and J. H. Hubschman. 1962. A comparison of larvae and larval development of species of eastern U. S. Palaemonetes with special reference to the development oi Palaemonetes intermedius Holthuis. The American Zoologist 2:394-395. . 1963. The larval development oi Palaemonetes kadiakensis M. J. Rathbun in the laboratory. Transactions of the American Microscopical Society 82:185-197. Bruce, A. J. 1993. Leptoplaemon gagudjui gen. nov., a new freshwaterpalaemoni d shrimp from Arnhem land, and a re-evaluation oi Palaemonetes holthuisi Strenth, with the designation of a new genus, Calathaemon. Hydrobiologia 257:73-94. Chace F. A., Jr. 1954. Two new subterranean shrimps (Decapoda:Caridea) from Florida and the West Indies, with a revised key to the American species. Journal ofthe Washington Academy of Sciences 44:318-324. Colless, D.H. 1970. The phenogram as an estimate of phylogeny. Systematic Zoology 19:352-362. 94 Collins, J. T. 1993. Systematic implications of allozyme variation in freshwater species oi Palaemonetes (Crustacea, Decapoda) from Texas and Mexico. Master's Thesis, Angelo State University. 44 pp. Creaser, E. P. 1932. The decapod crustaceans of Wisconsin Wisconsin Academy of Sciences, Arts, and Letters 21:321-22%. Creaser, E. P., and A. I. Ortenburger. 1933. The decapod crustaceans of Oklahoma. Publications ofthe University of Oklahoma Biological Survey 3:13-47. Digby, P. G. N., and R. A. Kempton. 1987. Multivariate Analysis of Ecological Communities. Chapman & Hall, London. Pp. 206. Dobkin, S. 1963. The larval development of Pa/ae/wowe/e5/?a/wflfe>5M5 (Gibbes, 1850) (Decapoda, Palaemonidae), reared in the laboratory. Cmstaceana 6:41-61. . 1965. Abbreviated larval development in caridean shrimps with the description ofthe larval stages of some species from south Florida. PhD. Dissertation, University of Miami. Pp. 228. . 1971. The larval development of Pa/ae/wowe/e5CW7w/wm^ Chace, 1954 (Decapoda, Palaemonidae), reared in the laboratory. Cmstaceana 20:285- 297. Farris, J. S. 1972. Estimating phylogenetic trees fromdistanc e matrices. The American Naturalist 106:645-668. Felsenstein, J. 1984. Distance methods for inferring phylogenies: A justification. Evolution 3%:\6-2A. Fleming, L. E. 1969. Use ofthe male external genitalic details as taxonomic characters in some species oi Palaemonetes (Decapoda, Palaemonidae). Proceedings ofthe Biological Society of Washington 82:443-452. Garcia, D. K. 1991. Population structure and patterns of speciation in the genus Palaemonetes. Doctoral dissertation, Texas A and M, College Station, Texas. 131 pp. Garcia, D. K., and S. K. Davis. 1994. Evidence for a mosaic hybrid zone in the grass shrimp Palaemonetes kadiakensis (Palaemonidae) as revealed by multiple genetic markers. Evolution 48:3 76-391. 95 Gibbs, L.R 1848. Catalogue ofthe fauna of South Carolina /w. Tuomey, M., Report on the Geology on the Fauna of South Carolina. Appendix, I-xxiv 1850. On the carcinological collections ofthe cabinets of Natural History ofthe United States: with an enumeration ofthe species contained therein, and description of new species. Proceedings ofthe third meeting ofthe American Association for the Advancement of Science 3:165-201. Heller, C. 1869. Zur naheren kenntiss der in den sussen gewassem de sudlichen Europa vorkommenden meerescustaceen. Zeitschrift fuer Wissenschaftliche Zoologie \9:\56-l62. Hobbs, H. H., Jr., H. H. Hobbs, m, and M. A. Danial. 1977. A review ofthe troglobitic decapod crustaceans ofthe Americas. Smithsonian Contributions to Zoology 244:l-\S3. Hobbs, H. H., m, and J. P. Jass. 1988. The crayfish and shrimps of Wisconsin. Milwaukee Public Museum, Milwaukee, Wisconsin. Pg. 114-121. Holthuis, L. B. 1949. Note on the species oi Palaemonetes (Crustacea: Decapoda) found in the United States of America. Proceedings ofthe Koninklijke Nederlandse Akademie van Wetenschappen 52:87-95. . 1950. Subfamily Palaemoninae. Part I in the Palaemonidae collected by the Siboga and Snellius Expeditions, with remarks on other species. Part X in the Decapoda ofthe Siboga Expedition. Monograph 39ct in Siboga Expeditie. 268 pp. . 1952. The subfamily Palaemoninae (Crustacea: Decapoda:Natantia) ofthe Americas, n. Alan Hancock Foundation Occasional Papers 12:1-396. Houck, M. A. 1992. Morphological variation in an ectoparasite: partitioning ecological and evolutionary influences. In: J. T. Sorensen, and R. Foottit (eds.). Ordinations in the Study of Morphology, Evolution, and Systematics of Insects: Applications and Quantitative Genetic Rationales. Elsevier, Amsterdam. Pg. 277-308. Houck, M. A., J. A. Gauthier, and R. E. Strauss. 1990. Allometric scaling in the earliest fossil bird, Archaeopteryx lithographica. Science 247:195-198. 96 Houck, M. A. and B. M. OConnor. 1998. Morphological variation in Hemisarcoptes (Acari: Hemisarcoptidae): Application of multivariate morphometric analyses. Annals ofthe Entomological Society of America 91: 335-349. Hubschman, J. H. 1974. The larval development oi Palaemonetes intermedius Hohhuis, 1949 (Decapoda, Palaemonidae) reared in the laboratory. Cmstaceana 26:89-103. Hutcheson, J. H., J. H. Oliver, Jr., M. A. Houck, and R. E. Strauss. 1995. Multivariate morphometric discrimination of nymphal and adult forms ofthe blacklegged tick, Ixodes scapularis (Acari: Ixodidae), a principal vector of the agent of lyme disease in eastern North America Journal of Medical Entomology 32:827-842. Jolicoeur, P. 1963. The multivariate generalization ofthe allometry equation. Biometrics 19:497-499. Kluge, A. G., and J. S. Farris. 1969. Quantitative phyletics and the evoultion of aimurans. Systematic Zoology 18:1-32. Lowe, B. T., and A. J. Provenzano, Jr. 1990. Survival and reproduction of Palaemonetes paludosus (Gibbs, 1850) (Decapoda: Palaemonidae) in saline water. Journal of Crustacean Biology 10:639-647. McCutchen, H. L. 1983. Morphological variation of appendicies masculinae in North American species oiiresYiWdXeT Palaemonetes (Crustacea, Decapoda). Master's Thesis, Angelo State University. 30 pp. Meehean, O. L. 1936. Notes on the freshwatershrim p Palaemonetes paludosua (Gibbes). Transactions ofthe American Microscopical Society 55:433-441. Neilsen, L. A. and J. B. Reynolds. 1977. Population characteristics of a freshwater shrimp, Palaemonetes kadiakensis Rathbua Transactions ofthe Missouri Academy of Science 10:44-57. Nelson, D. H. 1982. Thermal tolerance and preference ofthe freshwatershrim p Palaemonetes kadiakensis. Journal of Thermal Biology 7:183-187. Pereira, G. A. 1989. The cladistics, taxonomy, biogeography, and evolutionary history ofthe shrimp family Palaemonidae (Crustacea, Decapoda, Caridea). Doctoral Dissertation, University of Maryland. 418 pp. 97 Rathbun, M. J. 1902. Descriptions of new decapod crustaceans for the west coast of North America. Proceedings ofthe United States National Museum 24:885- 905. Say, T. 1918. An account ofthe crustaceans ofthe United States. Journal ofthe Natural Academy of Sciences of Philadelphia 1:57-458. Smalley, A. E. 1964. A new Palaemonetes from Mexico (Decapoda: Palaemonidae). Cmstaceana 6:229-232. Sollaud, E. 1923. Le developement larvaire des 'Talaemoninae. " Bulletin biologique de la France, et de la Belgique 57:509-603. Strauss, R. E. 1985. Evolutionary allometry and variation in body form in the South American catfish genus Corydoras (Callichthyidae). Systematic Zoology 34: 381-396. Strenth, N. E. 1976. A review ofthe systematics and zoogeography ofthe freshwater species oi Palaemonetes Heller fromNort h America (Crustacea: Decapoda). Smithsonian Contributions to Zoology 228:1-27. . 1991. The significance ofthe larval development oi Palaemonetes antromm Benedict (Crustacea, Decapoda). On the origin of subterranean palaemonids. Stygologia 6:\49-\53. . 1994. A new species oi Palaemonetes (Crustacea: Decapoda: Palaemonidae) from northeastern Mexico. Proceedings ofthe Biological Society of Washington 107:291-295. Strenth, N. E., and G. Longley. 1990. Reproductive patterns ofthe subterranean shrimp Palaemonetes antrorum Benedict (Decapoda, Palaemonidae) from central Texas. Stygologia 5:221-224. Ulenhuth, E. 1921. Observations on the distributions and habits ofthe blind Texan cave salamander, Typhiomoige rathbuni. Biological Bulletin 40:73-104. Vandel, A. 1965. Biospeology. Pergamon Press, New York. 524 pp. Villalobos, A. and H. H. Hobbs, Jr. 1974. Three new crustaceans from La Media Luna, San Luis Potosi, Mexico. Smithsonian Contributions to Zoology 174:1-18. 98 Webb, D. J. 1980. Electrophoretic systematics and population genetics ofthe genus Palaemonetes (Decapoda, Caridea) in Texas. Master's Thesis, University of Houston. 98 pp. Wiley, E.O. 1981. Phylogenetics; The Theory and Practice of Phylogentic Systematics. John Wiley and Sons, New York. 439 pp. Worth, S. G. 1908. Freshwater shrimp, a natural fish food Bulletin ofthe Bureau of Fisheries 28:853-858. Wright, 1954. The interpretation of multivariate systems. In: O. Kompthome, T. A. Bancroft, J. W. Gowen, and J. L. Lush, (eds.). Statistics and Mathematics in Biology. Iowa State University Press, Ames. Pg. 11-33. Xuan, N. V. 1997. On a new species of prawn, Palaemonetes camranhi n. sp. from brackish water of south Viet Nam (Decapoda:Caridea). Cmstaceana 70:85- 94. 99 APPENDIX 100 00 E T- <0 ^ CO o> •«- »". 1^-^. OT^^O)co^o o , CM 00 CO O SJ^2S2122*5<50)0) 3 •4.J CA ooa>ooa>oooor^ooa>(Oh~(0(0<0(0(Oooco<0(Dh^h~(0(Oh^(0 Vi 2 N c 3 o ed (0 (/) M CO CO CO CO CO CO td 3 3 3 3 3 3 3 3 3 >«>s>«>«>v>s>s>s 3 C C C C C C C C C CQCQCQCQCOCQCOCQ a. .9 .2 .2 .2 .2 .2 .2 .2 .2 (0 (S (0 CQ CQ CQ CQ CQ CQ COCOCOCOCOCOCOCO o a>o>o>o>o>n>o>o>o> X X X X X X X X X •o"0"a"0"0"0"0"a 333333333 0) cccccccc ed cu Q.Q.Q.Q.Q.Q.Q.Q.Q. 0) 0) (U 0) (D a> cu 0) T3 C cd Xfi CDCUO) CDCDCDO CD .2 i2 .2 ^ EEE-2.£ii.2ppppiJp.2.2.2iJ EEE — — — i^E — •I CD CD fsj S. S. S. ^ 101 Iir^!^^*S'^'^"*oo<*>cooou)CMO> "^ _ Vi cooog>ocorNjoo(Ocou)cncM^i^'<-cocoh«^co^iooococMa> ^'^o^cpooa>CM<}>iOT-cou)m^^oO'«-ooooT-cDr<^^oooo ^ O O O lO "^ 0-«-00<3>COCOOI^r^^r>-^»OI^ *5£2Tr^^'''**'^ »COCOCMr^OO-«-CMr^ CMCOCMi-CMCMCMCMCMCMCMCMCMCOCOCMCOCOCMCMCMCOCMCMCMCM co(Ocor^coi^cococo(or^cocor>^r^r^cDr^r«-r«>r^i^mincotf> H ID c o cocococococococococo 3 3 3 3 3 3 3 ._ ._ C C C cc >s>«COCQeOCQCOCQ 3 CQ CO "OOOOO O s cqcscQcocococococococococo u o 03333333333COCOCOCO o «5«>xxxxxxxxx cu C33333333330000 ;i:iEEEEEEEEE CCOCOCOCOCOCOCOCOCOCO.C.CiZ.C X 2.2.2.2.2.2 .2 ^^_a>^^ ^ o) C/3 EEE—EE—E — —E — — — EEEEE — — .2C.2.2C t3 D J3 3 u T-CMC0^cOC0r»-00 102 »'>^^'»-C0r«-C00>CM00O0)C000cr)C0lO<3>0>C0^IOCM0>C0^ :r^'~'*0>C0^C0CM'«-O»Oin00CMr^^T-|^JO00CMCDC0»OCM 00 E cj>CMcooo ^^COCM<3>r«-OOr^O>iOCMCOO>CMO)Oih~a>'^^COOOCMOOCMCO i^flOcooocom^ooT-iOincocococnT-ooococMOooooocoCM CM CO O (J> CO oC D 2 -c McooiOT-"^toooooijor^cocbd6hA ^^ • ^ ^_ ^^ ^^ 4M« «w^ a ^^ k^ ^^^ ^^^ ^A »*. ^_A . C«^D^ C«^O^ m• ^iik ^^, «^ • ^v • tfw^ ^^ ^fk frs f^\ r>-co(j)ooT-mcMr^<3>cocMoo^io O O ^ f- CO t- O to C>A KO to KD ed C oooo6da)oiho>o>o>a>a>r^dcM'^uScdco(OU)ir>od'^ir>in ed J Oh-^OOOOCM-<-^COCOCMCOQOCOCOCOtf)T-00(3>^COCDOOin coco^ocDooocooo C0CMOOT-"^cn»n00 CMCMCMCMCMCM"*CO"^CO"*'«fCOlO"*COOOOOOOOOOO H3 lOiOCDC0C0C0T-00OC3>O 1> a. a> a, c o COCOCOC0COCOCOCOC0CO^^^5^S = = E^ 3 -co -co -co CO -w -co E '£ '5 •£ S "5 '5 £ '5 i 2 E E E E E E E E E O ?l?l?>>>>>>>>>CQCOCQCQCQCQCQCQCOCa X CD CD u h-OOOOT-CMCO-^CDh-OOOOT-OICOr-CMCO^JOCOr-OOCJ)© 3 OOOT-t-i^T-T-'t-T-T-T-CMCMCMOiOiOiOlOiOmiOmcOCO O u)u)mininiomininu)inioir>inint^inminu>iou)iniou)u) o > CJ co"*iocoh*ooo>OT-cMcO"*»ncor*-oo 103 r^CMoocooo»ooO'^o>oocDO>eoT-T-coo)r^»oo>CMO>co»ocMr^ 00 5CM^oooT-oocMoooooooiOT-"^cO'-cDtf> oor^cooocMcomcoco OOCMT-CDCOOIO»OT—CO O CM ^ CO CMCOI^CMT-Or>^00»OCOOOCOCOCO"^0>flO"*"*cpijO1^ ^ o JC ^ O CO lO eud <0 lO CO CO t»-cocoioif> CM»O^0>h-.^CM^^C»OC0000000T-iOlOCMOpr^»OCM"J0J •* CM r^ r^ CM CO o "^CMiOmOCOCO"*OOCM"^ o)CM^iOT-o>mcoo) T- O O CM CO 0> r^"*cocOT-r^o>»n"^c3>cM '^mocou^oocooooo lO T- <^^ 0-^..^wh^r^ooincoio^T-ioc0 T- O o o>cor^^oooo>r*-io 'm, ^ o COOCMpppppp ^ P « lO 1^ ^ CM ^ p p "* P U> p « 1^ CM cd cd id o) d cd 1^ 1^' d CO k« od d d cd CT> cd ^" C3> od o" cd lo cd ^ r^ ^' CMCOCOCMCMCMCOCMCMCOffT-^^T- U)COCO>n-^CO^incOtf)CMCM-<-CMCMCMCMCMCM^CMCMCMCMCM(0 "5 « H3 CMT-ooor^OT-cMor«-cor>-h*r^r>-r»-r^t«~r^o>r^h-h-h*r*- c o iixiiiiii>>>>mmm I I I I I • I I I I I • I I I I •o"0"0"a-D-o"0"0'0"CD X CDCD(D(DCD 0) u •«-CMCO'*«n 104 ooT-h*cooo^cooo'^r^T-ooo>coocot^o>iooomcMcoco»o T-coocooioO'^CMtntncof^ocomm^oomOT-i^ocMcoco 00 G CJ •-r*oooor*-^m»ococMT-'^CMT-^oooocMO)coo>oO'«- P^-r-OCOO^T-OOCOOO^^CMCMlO^T-CMOOppT-p^ liiu i d ^' ^ CM ^ ^ ^ 1-" •r-^ T-^ ^ ^' -r-' ^ -r-' f-^ ^ T-" T-^ ^ -t-^ ^ ^ f-' -r-" ^ Vi u HViH moocomoocooor^O) OOCOCO r^"*0>00f^^^00^00C0l0OI^C0CMCMC000C0 OT-cMoooor^o ui n r^cooir)ooT-u)oocMcocotocMoocoi/>mooo>coo>ommo> h>-Or>^CMC0inCMT-C3>"*r-^OO"*^00OT-^CMT-cOCM00OiO ru m r»-'«-^ioior«*io^^coco^r^ CMCMCMCMCMCMCMCMCMCOCMCMCMCMCMCMCOCMCOCMCOCOCOCMCOCO Teet h Lowe r h-.h-OOCOr^COCOOOh>.t>-OOh-OOI»-l>^OOOOI^I^r«-COOOC3>OOC3>00 oa> cu c o O (!) O O O mmmi-i-i-i-i- S S _l _l _l CO CO CO CO Vi s 3 3 3 3 3 1 1 1 1 1 ' CO CO CO CO CO 1 .^ J£ .i£ .^ .^ .^ O CO CO CO ^ .^ .^ .^ .^ .^ o o o o 3 O) CO CO CO CO CO CQ CQ QJ CO CO CO CO CO CQ CO CO CQ •o T3 "O •a T3 cu T3 T3 •o •o T3 "O 3 3 3 3 3 o CO CO COCOCOCOCOCOCOCOCO CQ CO CO CO 00 CO CQ CO CQ CO CQ CO CO CO CO Cu .^ .^ .^ .^ .^ .^ CU Q. cu Q. Q. CDCD(D •T3 0^ x: »ncoi^cMco^iocof^ooo>OT-cMcO'«t40co^»oh*cor^ooo>o 3 o T-'<-T-U)U>U)iniOmiOU>COCOCOCOCOCOCOOOQOOOC3>(3><7>C3>0 c 3 CO(OCO(OCO(OCOCOCOCOCOCOCOCOCDCOCOCO(OCOCOCOCOCOCOI^ -tCi O O > CJ iocorx-ooc»OT-cMco »ocor*-ooo)OT-cMcO'^cncoh*ooo>o r.U G OOOOOT-T-T-T- ^•^•«-^-^CMCMCMCMCNJCMCMCMCMCMCO < E a> X> 4> ed CU H C« 105 or^c2i^'^<<-cj>oocMcoooooc3>rWJ BU t"M «*» «J «W OJ I^^ CM^cocMCMooooo>co»nr^tN •r- t»* 4M i>i ».»*».» w* vw "^ 1^ ow CM^cor»-^^o>'«-o>eoo>ooio»oiO'^oo^»n«£) »OCO WcOCMCOlOCMOO^meO^CM-t-T-CMOOCMinoOOO^OOiOOOT-T- CJ oof^oomo^-cocMcocotOh^ ^ cj..o C. O o_> .lOCMCO h«-COCOI^lO<3>OCMC3)lO'«-<3>"^COI^OOC3>lOCMOr^r*COOT-r^ E ^<»oioc3>h-iooo^O"* ru m 3 ^0>C0OC0<»"^C0"»-0>lOI^T-CM00lOC0000000r-^'»-"«-CMI^ •i-i ppph«.CMT-;ppr^ppcO^^CMCMpp-r^^^Pp'«-:PP uUl Vi ••-^did'^^idod^dr^cddcdodded'^CMidoduSod'^d'^oCM CM CM CM CM CM CO 'I*" CO CdO O CJ coco^coco^iococo^^^mco^cococococococoioioiO'^ J3 Ul CJ O H a>ooooooooocoh>oor^<30h**r*>'a>oa>ooo>o>a>coo>o>oa> ±1 a> t> cu a> cu c COCOCOCOCOCOCOCO o CO CO CO CO CO .3 .3 .2 .5 .2 .3 .3 .3 3 3 3 3 3 "O "D "D -a -O -O -a -O «_ i_ u i_ *S i5i555i555®©cD©cDa>a)a) .Q.QJ3.Q OOOOOcCCCCCCC '^ "^ -^ — '— — "^ .^ .^ O O rt rt 3 T3-o-a-o"oEfcEfcbttEc = t= = c: = ccc22SS cu -2.2=.3.3©©©©«©©©OO20OOOOOOOOO o (B cn CB (B (B^.^^^^'*'*-'*'*'*'^^^^^^^: SZ SS S: CO CO CO CD cu Q. Q. Q. Q. Q. .E .E .E .E .E .E .E .E S S S Z S S •£ Z S 2 5 2 2 CD CD CD (D CD CD CD a> 0) CD « CD CD CD 0) CO CO CD CO CO CO CQ t/3 CO CO CO % _ ^ fl^ <1^ <° CD 0> 0) CJ CD a> CD CD CJ 0) 2 LL u. U- U. LL u. LL u. 2 LL U- LL LL U- 2 LL LL •o u< ^ ^ a> CM CO •* in h- o CM CO "^ (O CO t^ OO Oi o CM CO -* lO CO »»- 3 ^m oo o> o C 3u o O o o o O ^ ^— T~ ^— T— ^— V— ^" T~ CM CM CM CM CM CM CM CM CM ro O 1^ r^ h- h~ r^ 1^ 1"- r^ h- h- h* h*. 1^ 1^ h- r>. h- h- 1^ 1^ r^ r^ »^ 1^ h- r^ G O > CJ CM co -* m co h- 00 106 o 1^ CM r«- 2)t*-OiOI^O^COO^OOOOOO 00 CO CO f- 00 E 22D;2?***'^<*'tf>'^'-»'>«ooocococo O 3 ^oooT-'«-cMinioioh«^a>co^m T- ^ CM in ^ CO CO o o ••-» JC CO o T-' d d d •-CHO 0) OCOCOT-IOCOOOCMCMCO o o o o o o <3> 00 -^ CO COO>^COO>h».^-OOCOOO o o o o o o ^ ^ 0> CO CJ JS lO O CM T- O O CM T- CJ) <3> Cd oo>ioo>ocMinp"^T-p-r-ppCMl~ O 1^ 00 CO ^ CO CU ^ CM CO CJ> •<- 00 ^ ed ^r^oddoiddcM "* CM r^ CO CM T- C3> d CM d d u CJc d ed HJ CJ ^ CO 00 CO t-eor^o^cocoio-^^oooooo 00 "* lo r»- E JS COO'^-OOT-OT-O'^X-OOOOOO pa>'^co^>r>CMa>ioa>cocO'r-cocooo <3) 1^ CD CM 2 ^ ^pcMp^pT-ppT-pcMior«-ioio 1^ T- T- p odindi^r^h-^r^d'T-^dcMCMddidid d CM d d Le n Ros t OCOOI^CMCOCOIOOIO o in ^ T- ui n i0iAC}>0>CMOCMO00h^ 00 00 r«- CM cocoooo-^r^oo^cMoo CD O C7> 00 pppr>-pp^^cooo CD ^ ^ P ru m •4-J CO uUl ^dt^ddc«idddd ai d d d O CJ CM^^T-CMCMCMCMCOCO COCOCOCOCOCOCOCO'd-COCMOCO^CO^ JZ u 0) fc o CJ o C Q CJ H ed i2 0?/33 CCOO <7\ OOOOOOOOOO'T-om'^CDCDCOCO rr U O CU 0) cu c o cccccccccc ._._._._ OOOOOOOOOO.-._2>2*2'2' 1 E E E E E E E E E E -52 .52 .E .E .E = 3 CDCDCDCDCDa)CDCDCDCD33EEEE cu o 222222222 — ^'^ P P P P cu C0C0COC0C0CDC0C0C0CO^^3333 Q.Q.Q.Q.Q.Q.Q.Q.Q.Q..C£: o u o o X CD CDCDCD(D CDCDCD V C/3 E^^I^E^EIE^IE^EI LLSS2SU.2LLU-LLSSLLSLLU. •o Um 4> ^ CM in CO -* CD V JS ,- CM CO '* in CD r- 00 Oi o 3 o3 CO CO CO CO CO CO CO CO CO -^ 1^ 1^ h- h- h- 1^ h- h- 523 9 523 9 in 060 8 c 022 7 r^ r«- r^ 436 1 "tcj o Oi o > '" '" •^ •^ •^ U r- oo a> o ^. CM CO -* in (O h- 00 Oi o ^_ CM ^ c in in m CO CD CO CO CD CO CD (O CD 107 00 p 1.194 6 1.498 2 1.443 1 1.377 3 1.786 3 1.484 7 1.072 4 1.078 6 1.007 9 1.049 0 1.177 3 1.328 1 0.880 8 1.129 0 1.775 6 1.263 4 0.975 2 0.967 7 1.345 5 1.400 2 1.510 8 .2.1 1.385 1 1.403 5 1.315 5 2.050 7 1-1 'G 2.033 6 00 ^ 883 5 806 4 719 3 660 3 662 1 844 0 000 5 198 7 753 5 766 2 980 6 115 3 122 6 542 0 093 3 898 2 548 2 017 3 740 3 891 7 014 6 453 6 .376 9 .903 1 .056 7 .732 4 ^- T— ^r" ^- '" '" o o ^~ o o o '" '" o '" "^ •^ '" '" o o o o '^ o CN ^ 00 S 189 6 167 7 706 6 243 0 040 0 388 2 520 7 855 8 016 2 304 1 463 3 934 3 808 1 811 4 989 6 804 6 252 6 203 8 210 7 462 1 446 4 .828 7 .503 8 J o .896 5 .750 5 .239 1 CO "* CO ^ CM CM ^ CM CM CO CO "^ CO CM CM CM CM CO CM CM ^ ^ ^~ CM CM ^ 127 6 179 7 162 0 500 0 527 3 444 3 730 5 985 6 922 3 869 8 016 6 933 1 376 0 010 1 295 5 651 0 403 9 072 9 090 4 226 1 647 4 417 4 643 7 .035 4 .305 4 .855 8 CO CO CO CM CM "* CO CO •* CO CO CO CO •* CO CO CM CM CM CO CO CM '" "^ '" •^ CN ^ 129 3 167 5 106 3 346 4 566 5 902 3 818 9 971 3 555 6 099 5 723 2 883 6 029 7 960 2 939 8 872 9 073 8 050 5 890 1 989 9 887 5 HJ 2 247 1 649 5 227 9 475 5 477 0 in CM CO T~ CM CM CO CO CM CM CM CM CM CM CM CM CM '" '^ •^ •^ ^ ^~ '" '" ^ ^ £ 163 4 521 2 581 0 872 5 709 2 562 7 778 4 776 9 555 8 857 7 048 4 753 7 334 5 527 7 781 6 504 1 569 0 533 8 922 7 436 9 648 2 081 5 998 7 228 0 650 7 692 9 o CM CM CO CO CM CO CM CM CM CM CM CM CM CM CM CM CM CM c^ ^ coincMcoincMinoocMoocMoo-«-cj>in"*CMCM"*incj)oocMCMco"* (Oooinh>CMincM'^CMcoina>ooincor>-cDCDincoincMCDC7>o 00 —. or^oot««-oin^i^cM'*in"^">«tcj>'«-cj>oooinoocD"*"*oo^in 4) >> c7>T-oincor«-incor*-o>or>-cj>i^oocooooococDinininin-<-in O'T-fOOOOOOO OOOOOOOOOOOOOT-O coooooT-oo'^r*-CM'«-cDcoo>r>--CMcoococMt-coooooo>r^'»-r^ eOT-inr^-«-'«-T-incocoincMO>coooo>OT-ocMOCMOinh-oo 00 3 cocMr^cM^cj>-«-r»-CMcooocMcoincM^in-«-ino)oo^ininh- r«-T-CJ>^CMCOT-'^^l^h«-COh~'*^COCOCOCMCM-»-T-0000 •«- CM CM -«- -H' CU ^ o o>^coor^co^T-in^cMinoocor^'*-t**t^r^CMCDoocMi^ino T-"^inc3>CMOooco^cDO'^ococMCM-«-incoooT-r^in^coco a> COincOCMCOCOT-OOCJ)CMCOOOCOCMCDOOCMh-CMOOCOT-T-^COO 3 CM"*^I^^CMppppp^^ph-;pCMp'*pT--«^0O0Oh*CM G cdcdcdr^CMCMCMCvicMcdinCvi^cdcMCMedcdcMCMCMCMT-^CMCMCM •^ o 'i ^ CJ o CJ ina>^ooocD^~in^C7>cooooooinr^a>a>ooc7>coc}>cor^^(D •«-incD^^-cOT-ininh-0)CMCMO>CMino>^-cocMinT-T-oocoin r^cDi^inooo>'«-CMCMr^oo^'«-ocO'^cMCDCMO>^oot^oooin < coinooincDi^cDoooO'«-^oppCM^coh-T-o>inco^ooo>h- -I 5 CMCMCM'f-^-r^T-^T-^'r-'^'cdcdCMcdcMCMCMCMCMCM^T-^T-^-r-.r^^^ 1 52 108 coooc3)o>oocoo>T-o>inoc3>cDr*-coo>coinr^cMCMco^'^o>in ^oooh-coo^-ooooincMoooococoinr^-T-^-coinooooo) 00 p OOCD00^CD^-0>C0^-in^0>0>0>^-CM0>C0l^<7)f^'^Or*-C0 COi-inO'«-CM^-OT-'r-OOOOT-OCO^CO^-COO'^'-inOC3JCO HJ .2 o -«- OJ^CMOcoin'^coinr^T-cMCD^ocooooocDoocMor^CMO ^^coocoo'^T-ooocMO'^coinincoinr^oooo^cooo ooeococO"*in^o<3>i^CMCM"*oococoi^ooococoooinr^^r*- oooT-cooooooooor^r^pooT-cMpT-cMr^oooocoooo ^ CJ d -o ci o^-ooooooooo o o o o o o o CN "^ pina>inoinr^a>r^cococoro^inT-cococD^cMCMoo^in^ 00 S COCDCO'«-^CJ)r^CO^COCMOOCDCO'^r>.OOOOOh-h-COCMOOCM'* «ncMCDcocMinoo^c»^^oooor^r--*cMincocOf-T-coT-cDCM hJ o PPPp^pOOr^C3>r.-CDCO^'<4-C]>OOCDT-OOCM-<-CMpp-<7 CMCMCMT-^CMCM-i-^T-^^^-r^-r^CMCMCMcd-^CMcd-r^CMCMCM't-^-r^CM c P r^coinT-co^cooococDCMomT-'^cR^cococo^T-oor^ooco coo>oocoooooor>-cDCM^'^CM'»-cooo>r^'»-ino)CDr*-h-'«-T- 00 CO ^c»cM-«-oc»ini^coh-cMoO'»-cor^'*OT-^i^^c3>i^'«-inr** CJ 3 ppCMCMpT^pp^pp^pCMr^cj>"*oor»-'d;oopCMpCM^ "^CM"^cdcdcdcMCMcdcMCMCMcdcdcdcd"^"cMcdc>icMCMcdcMCMcd T3 C ff CN CJ coocooincoin^inin'«-a>o>cooocDC3>a>ooooh>cDininr^co 00 CMOCOI^OOinCM-«-COCDI^OOCMC3>"r-CMCJ>OOr«-C3>CO"^'*C3>OCO 4> CO cDocooinT-T-ooT-i^cM'*r^cj)ococooinooocoh-oocj>in 2 •^p^ppf-r^pT-phN.'^'r^pincop'»-incDT-ocMco'*CM CM-r^CM'i-^-r^CMT-^T.^CM-r-^T-^-r-^CMT-^CMCMcdcMCM-r-^CMCMCM-r-^-r-'CM C CN cD'»-i^ooT-'^^coincDCMr^o>cMooinco^coo>^o>cor>-^co 00 CMCOCMCOO>OCMCM'«-ininC3>^CO(3>0>CO(3>CMCD(3>C3>COin^CM E h-r'-oocMoO't-inT-ocMCMOOin^^cMinrx-cocoococMco^ ppi^ppph^-pcMpp^^cMpoom'^h.-ocOT-coooooin cdCMCMCMT-^CMT-^-r^CM^T-^T-^CMCMCMCMcdCMCMCMCMCMCMT.^T-^CM nd L schi u CN t-H cocMCDflocDcooocM-«-ocMCMr«-r>-incMoocor^^inooo^cMO 00 •«-incor«-cDr^oc»cocococoincoinT-ocoooin^ooo^ooo V ?^ ^"^C0OCDOin'«t-OT-CM"*CMC»"*CM'«-Ol^r^^00I^^^CM HJ t5 i^ppi^pi^ppr^pppr^pppppi^ppph>h..iocD •*.J ed dddddddddddddddd-T-^ddddddddd CO Q •«- oocM-i-CMCRinoomr^comcRooh-CMooooincMooooincMOcjJco inocococMCMoo^r^T-CMi^cocMOT-coc»OT-c3)r*-T-incoco a> r^coo>coooincMooo>^T-t-o>coo>^ocMincooocDcoT-oco 3 OO-f-CMCDCOCOCMpppppT^l^pp^T-lv-CRCRCMOCROOcO C 'S CMCMcdcMCMCMCMT-^CMCMCMT^CMCMCMCMcdcMCMr-^T^CMcd-r^-r^CM G O CJ ^OCMCMOCJ>CMC0T-h-m0000"*"*C0T-00C»OC3)OCM00CJ>0> coooc}>coc7>ooocoooh>-flO^CMcor^o^O'^'^ooooooin^^ 00 CO 4) 3 OCMCJ>CO-«-COT-T-CM^CMCMCOCD"*CDOO"^in^COCOCOCOCMa) < cor<-^cj>o>'«-ooh-T-;h-h-^pr*CMppoocMcoooh-coh*h>.o CMT-^CMT-^^CM-r^^CMT-'^^CMT-^CMCMCMr^CMT-^T-^-r^CM^-r^CM i 109 CDO>^COO>0>CO'^COCOCOOOO'^'^T-CMCMOCMCMOCMI^O'«- 00 p o>o>coco^cM^^r-r^ino>r»-oinoor^incMO)CMO'^cMOcoco 00r>-00^inO)CM^inCMCD^— OO^CM^f^CO'-CMCM^-f^OO^O PpCMp^pO'^'^^COCOOCMCD^-OT-CO'^pT-;^pCM'r-; •2 I T- •r- -t-^ T-^ ^ T-^ CM T-^ T-^ T-^ ^' ,-• ^ ^ ,-: CM ^ ^' -r^ "r-" cd -r^ T-^ d ^' -r-^ l-l 'G f*^ *m^ CM r;g>oo'^CMooooT-co^o>o>'^^incMino>oooooT-cDT- 00 „ 00 ^<3>^o>^t-cDooinT-^-r*-inco^-cocM<3>i^ooh-o>coo>5 «> >^ ms-mcOcOT-co^o CMCMini^T-|«^COC3>T-CDCMCMcOr;P>0Op 1^ •*-' ^incoT-inin^r^opppppi^pi**: l^pppoOOCDOp T-^'r-^'T^T-^T-^^csiddddddT-ddd T3 ti d^ddddcMCMT^ CN ^ o>T-coo>cD^OT-r^or^ ocDr«-oor»-co^^oooocOT-^o^ mooo-T-oocM^h-oo^"^ T-coo>CMh*o>T-cor^ooino^— ooo 00 S ocDO>ooinooin^cMooc;> ^cor*-CMCDCMO>cD'«-oocoi^r^<0'^ J o i^T-p'T-pT-pr-.pppcMcOT-T^p^cM^ •1- 00 CM ^ O CM O ^CMT-^CMT-cMi^^>i^cdcd"^CMcd"^d ^ d ^ "t-^ c P ^CDCOCOC»C3>OTj-'<^'«»-OOOOr^O)CDCJ>inc0^r*-CMCMCO<3>CO OCMOOOCOOOcOCOCOCMCMO>CM'<-OOT-OCJ)CO"»-OCOOr>-T-"* oor-.h*r^oocj>CM^CM^"*T-i«^-«-cDC3)incocoooinoinT-^ pCMh>-ppT|;r^COOOCMOCO^COCOCMr»-^CMppppCMpr«>^ cdcdcMcdcMcdcdT-^T-CMCMCM-r-^CMCMcd-r-^CMCM-r-^-r^T-CMT-^CMT-^ C^ ^ OOh-.COOCMh>-00"*OOOr>*-r-CMOOOCDt-CDOOinO"*OOCO" CN "^ in h>.r«-cooo^i^o>"*'T-ocMinr^cMcO'^oococoooinr>-^^ co ^oocooocMco'*ooinr«-.T-"^inooinh«-inooocM^cocMh- ^ B CM^COC»CO"»-r>-CJ>OOCMCMO^CMCDCDininOCMCDCMincOOCCD O T-incj>int-cDor^oo^t-CMh-t-cocDcoinh-cO"*^<3>CM^'* CMCMr-^CMCMCM-^-r^T^CMCMCM^CMCMcd-r^T-^T-^T-^-r^T-^^T-^T-^T-^ c5 ^ cocor^ooom^cDoooocominh-oooh-^cooococMO^cRCM cr>0'c-cDcOT-^inooinT-c3>CMooin'^T-r>.cMooinc3>cocMooi- 00 --. CMr^^c3>cooinc}>ooincDCDOOc3>inino'^cMCMa>c3>T-.ocM a> >> I—I o h«;Pppr^l^ph^b«;l^h-;ppph^ppPPPpCD0Oinh-CO ^^ ti ddddddddddddddd-r^dddddddddd ;2 Q T-cD"*-»-i»-incj>oocMO-«-incor^cMO"^CMoor^cocor^r»-cocD CO 0OCDCOO>inh-in^CD^-T-COCOO>T-00O>O>O>T-CO0OCDO>0OCM 00 3 r^r^'<-coino>CMcocj>ococDr-incococj>CMCM' 0>OCDC0C00>^-00T— "^COT-OCMOOOOOr^OCDr*-I^CDO>COCD c3>cDCMoooc3>T-r*in h-m^mcMcO'^inooooocMcOT-cDh^CM -o r^cM^CMt^cRmincDcomooT- a> in'^in' O CJ OOOC3)T-in^O>OOr^OOT-0>l*^^OC3)^0000^00)T-CDCM 00 »:> c3>oc3>c»inoo^inooocO"*'^"*coh-T-inoinincj>cj>T-'^co 4) 3 incoinooc3>c3>cj>co^coc3>cooocDr^ocDCJ>r^r^oooocoh>-cMCM < OOCOOOCOC3>COT-CJ>000)OOC3>^T-^-«-COinr^^CM^O-r-'^CO 52 •»-CM'»-CM-«-CMC0^T--«-^T-T-CMCMCO CM 110 r^O'»-o^-r^r^'^^oocoincocooco^o>^cooocMooo>ooco 00 p CMCMI^Oh-OOCOCDOOh>-OOCJ>COOOCO'«-'^-«-COOOO"»-CMinCJ>r^ 4> E CMcocoincMininmoooocoT-ocMor^oooT-mcoT-inoocD h-;^pppincMCD^i^o>T-o>ooooO''-o>ococo^^«nco •r^- ^h>cooocO'^ooinoincMa>coooo>cO'T-oooin^a>ooinooo '«-^o>^h-'*-coin^-o>ooT-^inooi^incooor^^or^cM'^^ 00 ^ h;0>T-ooo>cocDi^ooo>ocMOCMr^o>r^ooooooincMr^cD^h- l_J -*^ pp^cM'^mi^^cooi^ooh'-cooor^coh-cDooooooopi^ T3 ti cMCMcvicMCMCMcdcMCMcdddddddddddT-T^dddd CN ^ ooa>a>cO'T-'T-r«.o>^ooi^'T-ocDin^o>h>^i^o>oocooo^o CMr^coT-cDcococoT-coint^oooocoin'«-oo^cMinr*-^r^o> 00 S ooa>oooooocor>-a>oooincMCM^CMCMT-a>r^ooh-cD'<-^^CM J o ooo^T-co^co^cocomoo^^oi^^r^cor^cOT-coooor^ ^'^"^"^^"^cD'^'^in CM CM CM CM ^ T- p p <^ cu ooincocMcooooino^CMcocDT-r^coco^c3)OCMh-ininoo coooc3>t-a>oocj>^i-o-^in^c7>ina>incocDcoinoocD'^r^oo ^-co-T-^incMinT-eoin-^r^coocoinoinococMOCRoooco ppi^ph-ppi^^^cMi^comcomcor-^p^pppcMCM CMcdCMCMCMCMcdcMCMcdcMCMCMCMCMCMCMCMCMCMcdcdcdcdcd-r^ C«N CJ^ CM^h-incj>"*incooocMCMcoo>ocoocD"*c»"<*-c3>^^inincM i^cocococj>r«-r>-cDincooooi^cMooc3>'^oincococj>oor»-c3>oo 00 „ ocMoocoT-cj>oor-cMco-«-T-r^ooincor^cMininocDor»-in cMh*cj>oocMCMr^ocoino^ocoincj>cj>ocMcoh-ino>h-oin 00 —. inoinro-oocoinh-.coooT-comincMr-^r^^h-inT-ooocM 00-«-ininCDCMCM"*-«-CM"*T-CM^flOOOC»'^-<-^l>».00'^OOCJ>CO CM^pcMpph>-pppppr>>pppppppoor^a>copin .^J ti T-^ T^ 1-^ T-' T-^ ^" f-^ 1-^ T-^ -r-' d d d d d d d d d d d d d d d d cocMOOO>oocOT-'^cDO>cMor*-incor»-i^'^incocoooi^ocM CO OCJ>r«-'*COCO"*CJ>OCM^C3>COOOOOO'*C3>COh*OO^CD^CO 00 3 T-inooor*-oor*-cocoo)^oocj>coino>cj>cocooooocMinr^ooco •r-cOCM'r-CMT-0>'^CO^OpT-T-'r^O>OT-OOin^inCMCMO ^ a CMCMCMCMCMCMCMCMCMCM-r-'d^-r-'ir-'di-^"r^'i-"'r^-r^^.r^,-^.r^.r^ CO O h*CMCJ>CMCr>inO"^^OT-OOCMOOCMCDC3)CJ>l*^-«-CDCJ>OOCMCDCO a> CM comcMT-r^incocor^oocoi^CMooT-^rs-r^iocD'^oo r^ CO 00 CM CO ococMOco-t-T-coincDCJ>cocD"^r»-T-inT-cD'r-T- 0) a> h» 00 a> o oppi^pi^^ppppi^h-^ppi^T-T-in^co 3 CMCMCMT-^T-CMT-^CMCMCMCMCMr-^-r-^^CsicMcdcdCMCMCM ^G '•^ G sa O CJ omooooocDCMCMCMinh-O'^"^ cDCMCMcocoh-oooo>r>-inr>. oooino"*h-o CM^-'»-cj>inoo^h-oocj>'^"^T- 111 CMinoocMcooocoinooo>oo^coT-^r^^incoinT-incoo>o>5 cor^r^cocO'»-^cD^o>o>ooco^o>^-CMO>r^cM'«-o^cM'»-co inooo>^incoincM'^^-CMO>o>t-t-oo'^r^o>"*r^o)i^ooo iu m Le g 00>OOCDO>OCMOCOCOinOT-00)T-COOCMO'»-COin^pCM T3 'r^ddr^dr-^y-^T^T^f-^T^-r-^T^T-^d-r^-r^-r^-r^y--^-^-^-^'^'^ Ul sc h m I-H CMO>0'^0>OO^OCMOOCMCO^COOT-00^'«-'«-O^CMO>^0 0'^r*~o>CM'»-^-ocDOh«-CMh^ooocooor^coinoo^coinr^ 00 oooin^ocomococMcooococoi^cor^OT-T-h-cMoor^in >^ i^ppcooor»-oocooocMr^cooor»-h«-o>r«-h"f^ooo>h^ppCM HaJ u ddd-r^ddddT^r-^T^dddddddddddT-T-r-T- T3 ti C CN Q ^h~^cooo^^oocor^coc»co^incMOcoooin^co^r^cM CO cj>cocoooooT-cocO'«-incM"«-ini^ooooh-incoinc3>r- oocDOc»cococor^CMh--«-'«-i-coc3)r*-incM"^T-T-^oococD h-coooo>^-^i^oinooincMCMOT-r^oincj>o>ooh*cooo 00 in^or>-cDin^oooococMi^cMinoocDh-cj>co^oocM^oocMoo in^cDr^i^in-<-coincMooooooc3>r^cocoi^a>cDC7)coooo'^'^ 2 CM CM CO CM CM CM •T3 0) ^•«-T-CM"T--«-CM^CMCMCM P CM OOT-COCM-^T-lx-r^CO-^COCR^CDOOT-COI^OOCRinOCO-t-CMI^ 00 "^OCM-«-OO^OOh--T-^h-COCMCM^C3>CD^T-r«-"«tOOCOI^CDOO E inc3>inin'<-a>oococM(Ococo^cDoor<>CMCMoo^oo-<-oooinin OCJ>OOCOCOOCOO^I>-CMT-^r^r^T-OOCMCOOOT-CJ>"*CM'«-'«- CM •«- "^CMCMCMCMCOCMCOCMCMCM'X-CMCMCMCMT-CMT-CMCMCMCM nd L schi u CN I—4 r«*cDoo^cococj>r<-coincocD"*cDC3>ino>cocMOinh~"*incoo oincoc3>ooT-inr»-cj>ocMr^cooocMcoc»^coflOCMoocMcooo 00 r*-coincMr*-o>^coo>o>ooocOT-'»-eoinincoh>-i^cDOinor*- 0) >. ^in^c3>h^inr>-corx>-coooinincMCDCDcoincoincDi^or»oop HJ t5 ddddddddddddd-r^dddddddd-r^ddd *J ti CO Q cDoococo^CMi^oc3>cDOinooin^cocMcoint^r>-i^incMCMco Vi inc}>oocooocoo^in^oa>inoo>cMoooooocMCMflot^c3>h»cM 00 3 cDr-h-or^^cj>CMr^coco^r^o>^cMCM^oocD^^incoinr^ V T3 cj>cj>ooppppp^cMppppo T-cM-«-f-oco^oocoin~" ~ ^ HJ O ddd-r^-r^-r^r^r-^-r^-r^-r^-r^dd >J CU CO O •cor^r*-inr^cOT-coo)oooo>cDO)^ocMcooco^T-'^cD'«- • lOinT-r^coco^incor^-i^ocoo^T-ininr^i^oco^'^oo cor^-co^Oh-ooincoinT-oococMcoocoh-h-T-ooocooocOT- I^O^CO^-pppCMppppppppCM^p^COf-CMCO^ i -•i o ^•r-^T-^cdcMT-^CMT-^cdcMcdCNicMCMCMCMCMCMCMCMCMCMcdcMCMCM O c»CMCMCMoo^co 112 o>^^oocoincMr^^-o>ocDT-co^-coo^ocococooo>oco t^ocD'-cocoocDO^r^^in'»-oinoor^o>incMT-i^ino>o 00 p ^-CMin^ocMr^^ino>inocDr^cDincDcocMr^'«-r^'^f** ooo>CMinino>r*-cMcoo>ocDcoocoo^r^o>r^coinr-^cMO ^cocoo>^^-^ooino)o>i^cocMcooh^ooo>oof*-^"0>o>in 00 „ ininr*-^ococoo)coo>cD'^o>oocooocoo'^i^'-oocMcocor^ «> >. •^r^cj>h»inooo>coo>oocoo)cocooooflOO>o>^o>^cOT-cDco T3 ti OOT-OOOOO^T-OOOO^O^^COT- CN "^ ^ininoh^cD^oinh-^incDCMcomoocoocDincMh-^CMO CM^r^inco^»^inh*-ocOT-h-inoococM^oocOT-oocoooo 00 S r*-^'r-cDCMinooo>cooocMT-cOT-incoooocDcoinoo^of^^ ^pppph-pp'r-o^oincMCMO>o>oop'^pcM^ppp HJ O CNi"^^fdcd'«-^-r-"^CMCM^CM^'cdcMT^"»-^^-«-^cd'«-^cd^dcd"^ p p OCDCOCD^COOh--CMT-CMCMin"^CDO"*COCM^CMCD"«-^C3)C3> cj)0"*cDOCMooinincMcoo>cocoo>r^cocj>c3>OT-cj>ooocM^ SP '^ CMCDcOin-<-CM-^CDO>COinOC3>^OOh~OI------^ ^ooo-r-oomcDcD p^pp^ppcMppcoincor«-c3)inin inr^cococoor^p CMcdcdcdcOT^-r-cdT-^CMT-^-r-^-r-^CMT-T-^T-^ T-: T-" -r-' CM in d ^ "^ C^ ^ •«-CJ>'«-^COin" CO CM-^incMcO'^^oooT-T-incoi^m^coooh-coococO'^oco CMCJ>CDC0inT-CMCMC0inOC0h«.C0C0C0CDC000O^CM^h-C3>O 00 3 o^^ooin^QOcoinoooo>CMooh^ino^cDincDoo^T-^QO inoocj>h-i^OT-incM^r^^cMincMf-oooino'^<3>cor^h- ^ a CM CO O 1—1 Ul coincj>CMi^cMoocoinh*-«-cj)0-«-coincMo>cocDCJ>CMOh^inc3> T-h-T-C3)r«-'*CDCMr^CDOCMCJ>OOOOC3>'«-C3>T-OOOOCOT-^ ooo^T-cDoocoo>h-h-r^h-h*0'»-r*-o>cM'r-cD'^cMinooco 3 CM^^"^pp^^pppp^'*'^p^pcMpp^*cM^-lnocM P CMcdcdcdCM-r^^cdT^T-'d^T-CMCMT-T-^T-^T-^-r^T-^CMcd^cdcd ;•2^ uo o CJ ininoc3>oocor*-^^- 113 o>ooinc3>^-oinr^a>r«> O O O O O O O O CO O O O O O 00 E CMOO-^r^CMI^COOCOCM OO <3) in r^ C3> CO O CD ^ o E 0>^CMCMOC0'»-CM^00 CD 00 CM eo CO COO>OCMCM'«-^-incD'^ o ^ CM m r^d-r^T^T^y-^T^'r^-r^T^ CO CO o o T^ d d d ^CMCMOmOOOJCM^OOOOOO CO o ^ r^ 00 « r^ CO T- ^ cn -^ '^ 00 l-J •*-» CMpro-oooooooocococot^r^^cMT-p T- in o p T- o"ddd d o 1- o O T- T- -r- T- •—I o T^ d d d TS ti CN "^ CM^inocomocMinoooooooo 00 m ^ in inh>o>Qoooa>a>cooo>oooooo "5t <3> 00 ^ 00 S inh-OOCMT-'^COT-CO^OCMOOOO> CO m 00 r^ hJ o •»-;ph-CMpCM^pp^pCM'«-r-^p in •«- p T-; cd^T-^CM^CMCMcdcdcdT-^T-^cdCMCMCM CM ^' d d p p oocooococoocooooooooo "^ oo 00 r^ mcOCMCDOOCDCOinOCOOOOOOO in T- oo in SP «« OOOCOCD^OOI^COCOOO^r^CM^CO m o CD CO CMppp^ro^^pppr^cMCMr-^co p p p T- S 3 cdcMT-^CMCMCMCMcdcMcdcMCM^cdcdcd CM d d d c^ ^ CMOm'cl-COI^CR^OCROOOOOO O oo O CO C0CD^I^C3>inCM'*-«-"*OOOOOO CM r*- 1^ CM 00 ^ T-in^in0000C7>0>T-CMCDC000OOC0 CO ^ m ^ HJ 2 pinincj)"^-«-cj>c3>ooT-cj>r^T-"^CM'* CM r«- o T- CMT-^-r^^-r^CM^CMCMcd-r^T-^cdCMCMCM CM d d d CN "^ oh-r^-cMcomh-^ocooooooo (3> O O CD 00 p C3>cooin^ooinoooooooo in CM C3> CD r>-CMoocoinr^ooinh»oc3>cDC7>cocMoo CO "* "* CJ> O P cO"*coooooo>coocMcoooo^r^r>.o CM CD O O Jj 3 CMT-^T-^-I-^T-^I-: CM CM CM CM CO CM CM CO CM d d d c5 ^ CMinCMCM^mcD^COh-OOOOOO -* 1^ r^ C3> r^CDI^^OOCOCO^"^CMOOOOOO o 1- r- -^ 00 —. ^inT-coooc3>cocoincMCJ>T-|x-cj>cj>oo h- CO T- CO h- CM O O I^COCOCOCMCOCDOOOOOincDOOOOOC3> >j ti dddd-r^ddd-r^ddd-r^ddd dddd CMh*CDinCDCJ>^T-CD"*000000 o CO 00 a> CO C0-r-h>.C3>CDC0inOOC3>OOOOOO CM f- C3> 00 O C3> CM in ^^oococMCMCDcooco^in-^^mh- ^ CO o o inCM^CMr^COCMCOOI^CJ>OC»COCDOO 00 3 T^ d d d -r-^f-^T^T-'dT^f-^'r-CMT^dT-^'r^T-^T^T-" to O cM'^mcD'^ooT-'f-incDOOOOOO CO CD O CD OCMCDinO^CM'«-h-'*000000 CD oo T- o ocooococDh-^oo^oinT-incot*-in 00 CD in o ^CDCOOC7>CMOOOCMC7>'^CMCMppp p p o T- CM^T-^CMT-^CMCMCMcdcMCMCMcdcMCMCM CM d d d o U cocoT-coom^oocDOOooooo h- CM CO in o>CMcocoooocMa>ininoooooo CM 00 o CJ> 00 CO CMoominmcDi^cocMOCMOOT-cMoo -«- CM -* h- 114 ooco^-'«-inr^o>o>coT-^-ino>oincMoocM'^^o>inincMCMin ooo>r^coo>r^cDCMOcoococoooh*o>o>^o>ooinT-r^incMO 00 r*;P'-^CMincoc3>ooT-oococMO>r«-^cococ3)OOinO'«-o CJ E com^OT-ooooooT-coin^^^inmcocOT-oocM^CM 3 »0 HH h-^T-ino'«-r^oooococM'*CMCMOcooT-r^T-r^oo^h-o S)!!dJ'^^*"***>"*o>c»r^"^CMcoh-^coo>cMOcocMr^cM^r^ 00 —I ppincp^'^CMincoi^-^coocD^ococDOcDoo^incor^^ -1 tj ppT-|pCMpOCMT-C0l*»-0>C0CMCMCMCMC0'^O'^00r-;I^Pp ^•'-••'-•di-^d^"T-^T-^-r^T-^dT-^T-^T-^T-^-r-f-^T--r^dddddd 4;? Q I^OCOOOOOOI^COh^T-OOinoOCOmOCMOOOOOOCO^CMOOO CO ina>cOf-oincooocDcoh>r^h>CMr^h>incoooocDO^cDocM 00 3 incM^oincMcocD^-oincMinT-h-cM'*^coo>inoocMO>T-T- J O •r-ppp.r-h-.^C0CMC0h~OC000T-CMin00C0h-CMp'*-;PP'^ cdcdcd^CMCMCMCMCM^'^'cd-^cdcdcdcdcdCMCMCMCMCMCMCMCM 4;? 2 oO'^0)cocMinr^coooh^T-oo^oooT-h~^'^o>c3>o>'r-coin 00 CO coh>.ino' T-c3>cDO<4-CMOCMoO'r-coa>'<-cDoo^h>-coa>coincoc3>h-ino •«-ocO'^^i^c3>c»or^r^o^co^i^OT-cM^h-CMincDCMoo c3>oo-v-'^incocMi^inr^CMoooo<3>a>coc3>o-<-inooc7>^cDO^ p"^"^pcOCMCM'*CO-«-'^CDCJ>-T-0O0OCOCO0Or^^CMCMin0O"* |2 cdcdcd-i-^CMCMCMCMCM^^CMcdcdCMCMcdcdCMCMCMCMCMCMCMCM •»-CM0>C0CM'*000>r*-0>r^CJ>CDOC0CMC0CMr»-^O'*CDr*-00CD incMinoo^inr-incoooco-T-oooT-^CMr^^mooooooooinT- 00 E oocDC»ooT-'r-r*.coh-cooT-T-o>inr«>-h-o>ococM^cD" 00 — 177 4 131 3 337 4 531 6 879 9 959 5 755 8 563 6 920 1 276 0 064 0 914 8 980 4 078 0 066 2 100 5 450 5 626 6 470 7 285 6 931 6 747 1 802 1 826 2 732 8 878 9 ^ ^ ^ o o o o o o o o o '" o 639 3 1.03 9 CO 00 3 159 7 707 9 532 1 768 6 826 0 865 6 821 3 377 8 864 5 734 8 825 3 755 5 746 7 268 5 621 1 925 4 217 8 634 5 201 0 835 6 743 9 080 2 936 2 427 0 CM CM CO CO -* CM CO CM CM CM CM CO CM CM CM CM J O '" •^ •^ •^ 0)0 C3O 1.466 2 1.213 5 1.800 6 1.642 4 1.364 5 1.606 5 1.389 6 1.723 9 1.329 8 1.210 3 1.004 2 1.215 7 1.153 2 1.103 4 1.559 9 1.753 3 1.304 2 1.455 2 0.998 6 0.892 9 2.260 4 1.209 8 0.869 9 0.989 9 1.277 1 § •a ti 1.159 0 o CJ m^ 139 3 113 3 132 1 571 1 151 8 104 0 792 7 513 8 458 0 085 4 215 4 238 2 860 2 236 5 425 2 027 3 393 4 428 3 608 6 201 9 .574 3 .909 5 .888 0 .722 3 .112 8 .225 3 HJ 2 CO CO CO CM CM CM CM CM CO "* CM CO CM CM CM CO CO CM CM CM CM CM CM CM CM 1 115 iu m Le g 1.379 4 1.231 9 1.103 6 1.536 3 1.446 9 1.081 6 1.611 9 1.046 6 1.331 3 1.191 9 1.324 9 1.383 6 1.244 3 1.254 8 1.226 4 1.516 9 1.524 6 1.121 2 1.005 5 1.488 9 1.180 3 1.458 2 1.474 8 1.177 7 1.087 2 ,c 0.963 3 •«-» Vi m 00 990 6 071 5 950 5 p^ 804 6 060 9 135 3 818 9 028 4 708 6 914 9 153 2 095 5 787 5 926 5 092 0 400 3 656 8 732 8 252 9 253 3 a> .646 9 .092 1 •4-» .740 5 .029 4 .834 3 .987 9 J U .c ti ^ o '" o '"" •*" o o o o o o o o o o o o Q CO 00 3 TS 128 3 176 6 136 7 401 8 043 3 432 4 087 4 937 5 567 8 686 7 014 5 957 2 087 9 695 9 430 9 623 5 636 5 293 9 404 5 .185 0 .521 4 .170 6 .445 7 a> .962 9 .425 1 .584 7 U O CU CM CM ^ CO CO CM CM CM CO CM CM CM CM CM CO CO CO CM CM CM CM CM CO CM CM CO O TT u 0^ 00 CO a> 3 1.558 4 1.427 5 1.227 3 1.682 3 1.283 1 1.349 0 1.131 8 1.663 9 1.186 4 1.839 9 1.593 5 1.469 3 1.368 2 1.747 9 1.190 5 1.337 3 U 1.285 5 1.219 7 1.406 4 1.244 8 1.230 1 1.410 8 1.224 2 1.610 6 1.379 0 0.969 1 4S tiB - u 00 M 076 0 400 4 347 6 376 4 733 0 589 4 954 1 512 7 518 7 320 8 368 5 569 2 584 0 096 3 551 3 865 6 299 5 469 3 792 9 993 0 249 1 424 7 436 5 00 iu m 1.357 6 1.353 5 1.215 5 1.226 7 1.244 0 1.096 3 1.133 6 1.445 4 1.455 8 1.187 4 1.273 0 1.188 6 1.060 1 1.389 1 1.125 9 1.504 0 1.484 6 1.373 3 1.461 9 1.452 3 1.233 1 1.549 9 1.194 4 1.336 3 1.070 0 ^ 0.849 6 .«-* M "^ 00 179 7 150 5 944 4 528 1 362 1 724 7 747 5 808 5 925 5 914 3 202 1 987 2 919 2 746 1 308 9 017 1 787 5 092 8 808 5 775 2 042 8 a> >> 820 2 999 4 716 8 089 8 692 2 HJ tj o o •^ o •^ o o o o o o o o ^' ^ ^ ^" o o o •^ o o •^ TJ ti '" '^ Ul o CO 00 3 108 0 538 4 563 5 024 5 072 5 208 2 054 6 007 3 118 7 909 4 075 0 087 8 427 6 281 0 017 9 033 9 734 2 363 9 307 5 634 6 030 8 977 8 336 6 a> •o 353 4 915 7 213 5 J o CM CO CM CM CM CM CM CM CM CM CM CM CM CO CM CM CM CM CM CM CM CM cu '" •^ '^" •^ t3 o ro 00 CO 4) 3 3 1.372 6 1.342 9 1.396 7 1.132 0 1.525 6 1.084 3 1.191 9 1.044 9 1.041 3 1.028 0 1.069 8 1.050 1 1.258 6 1.207 9 1.142 1 1.147 1 0.900 1 1.265 0 1.220 8 1.140 0 1.125 3 1.631 9 1.076 2 1.417 0 _C 1.688 0 1.058 1 •o ti G O CJ 120 0 125 1 166 1 155 4 122 4 809 4 560 8 137 6 116 4 195 6 488 8 355 2 209 9 620 3 677 4 016 6 818 2 293 3 506 1 .937 3 .371 4 .439 8 .835 2 .183 5 .377 2 .718 2 J 2 CM CM CO CM CM CM CM CM CM CM CM '" CM CM CM CM CO CM CM CM CM CM CM CM CM CM 116 00 E 1.222 3 1.181 8 1.205 8 1.388 4 1.412 3 1.807 8 1.136 6 1.439 7 1.245 1 1.296 3 1.128 5 1.679 1 1.334 4 1.457 8 1.519 0 1.273 1 1.139 7 1.300 2 1.389 5 1.204 3 1.807 6 2.445 2 1.413 6 1.395 3 1.656 0 •21 0.992 1 00 — . 158 0 107 0 537 7 762 7 146 3 199 5 571 6 406 2 533 2 038 1 093 6 516 7 862 9 332 5 622 1 272 0 .524 8 HJ t3 .829 7 .000 9 .830 0 ,257 6 ,515 9 .533 2 .657 1 .533 4 .545 2 o T— o o T— ^- d o o o o o o o o o o 'G ^ ''" •^ •^ •*" '" ^ '" ^ •^ ^ Q CM 00 3 188 7 197 8 115 8 146 2 089 7 044 4 779 8 640 4 272 6 755 7 886 2 897 2 892 3 650 4 666 9 00 CO a> 3 1.793 6 1.567 1 1.444 1 1.503 4 1.769 9 1.265 0 1.441 5 1.563 4 1.573 8 1.506 2 1.369 2 1.832 5 1.259 0 1.855 7 1.965 2 1.619 9 1.510 1 1.401 8 1.501 9 1.093 7 1.653 1 2.403 0 2.150 7 J g. 2.629 0 2.400 7 2.313 8 ^ ti ^ o CM CO ?P c« CD CJ ^ 160 0 832 6 236 7 048 0 670 9 CO .541 7 .010 7 ,458 0 ,054 8 ,448 1 ,842 6 ,916 5 ,021 2 ,115 7 ,152 6 ,860 5 ,313 9 ,506 5 ,754 6 ,599 2 ,071 2 .274 8 .873 5 .375 2 .265 3 HJ 2 CM CO CM CO CM CM -* CO CM CO CO CO CM CO CO -* CM CN CM CM CM CM CO CM CM |2 ^~ 175 7 184 9 138 9 557 5 325 2 821 3 275 6 436 3 980 2 243 4 205 0 407 3 460 7 611 6 ,066 8 ,372 1 ,184 2 ,218 1 ,562 6 ,302 7 ,309 5 ,991 3 ,198 3 ,347 9 ,450 4 ,731 0 hLe g chiu m i: CO "^ HH 993 0 81.3 9 00 — (U >> 493 0 ,901 6 ,967 8 ,777 9 ,255 7 .809 4 ,276 6 ,712 6 ,855 6 .998 5 .224 8 .961 1 .999 1 .192 1 .267 1 .357 1 .491 3 .542 2 ,620 6 ,524 0 ,452 7 ,544 7 ,622 3 ,575 2 HJ tj o o o o o o o o T— o o o o o o o o o o •o S$ '^ ^" ^ '" '" ^~ '" ro '^ CO 00 3 586 9 585 9 501 6 053 4 797 0 T3 00 CO a> a> 3 1.432 7 1.793 7 1.250 5 1.424 6 1.271 6 1.493 1 1.745 5 1.454 0 1.169 7 1.398 9 1.805 5 0.921 7 1.216 5 1.704 8 1.134 4 1.479 4 2.052 7 1.965 8 1.366 3 1.383 4 1.217 1 1.196 9 1.078 5 2.440 6 1.517 4 1.288 5 Carp i 3rd L ontin u CJ ^ SP «« 960 9 381 1 812 5 754 5 674 2 559 3 088 8 033 1 268 7 870 9 666 5 < ,033 8 ,711 5 .852 8 .870 3 .504 2 .918 2 .038 3 .310 9 .385 8 .800 9 .023 9 .321 1 .009 6 ,896 6 ,157 0 J 2 CM CM CM CM CM CM "* CO CM CM CM CM CO CO •^ CM CM CM CM CM CM abl e ro '^ H 117 iu m Le g 1.341 5 1.690 6 1.609 8 1.489 8 1.594 8 1.314 1 1.654 3 1.565 9 1.763 3 1.592 1 2.087 4 1.208 1 2.014 3 1.683 7 1.980 9 1.045 9 1.154 6 2.253 6 1.115 0 1.029 3 1.108 8 1.090 4 1.064 0 1.074 9 J:; 2.208 4 0.985 7 ir> CO MM 00 a> >% 1.136 9 1.261 8 1.086 8 1.079 8 1.269 2 1.229 0 1.301 5 1.072 3 1.273 2 1.192 1 1.430 2 0.946 1 1.610 2 1.141 8 1.152 2 1.172 0 1.225 6 0.923 4 0.805 5 0.726 3 0.800 2 0.814 4 J 0.990 6 0.818 0 0.966 6 0.942 1 ^ ti Q CO 00 3 T3 110 5 154 8 148 5 139 7 088 6 967 8 330 8 384 9 850 0 073 9 633 7 265 3 a> 474 5 672 8 H^ .673 3 .603 1 .151 4 ,082 4 ,220 4 ,218 9 ,036 2 .232 0 .192 3 .145 9 .161 4 .744 9 O CM CO CO CO CO CO "* CO CO CO CM CM eg CM eg CM CM CO CO CO CO eg CM CU •^ •^ •*" •4-> O "^ Um CU 00 CO a> 3 ^ 1.208 5 1.373 0 1.570 6 1.868 5 1.698 3 1.876 2 1.519 7 1.596 0 1.630 4 1.699 9 1.372 8 1.576 0 1.525 7 1.780 8 1.749 2 1.904 6 1.070 6 1.346 0 1.102 7 1.156 8 1.232 3 1.058 2 1.031 9 2.026 2 0.899 1 0.994 0 CJ SP «> 122 3 105 8 817 7 389 1 088 5 552 0 069 2 076 3 056 1 479 3 242 8 449 5 MJ 2 ,717 8 ,826 6 ,829 3 ,647 1 ,031 8 ,725 4 ,038 8 ,167 6 ,157 1 ,294 4 ,369 3 ,928 3 ,440 2 ,433 5 CM CM CM eg CO CM CO CO CM CO CM eg CM eg eg CM '" eg eg CM CO CO CO eg eg eg 00 E 1.734 2 1.787 5 1.747 0 1.857 8 1.798 4 1.781 8 1.982 5 1.701 2 1.047 7 1.087 9 1.075 3 1.038 0 1.196 6 1.069 8 1.094 0 1.279 0 1.669 6 1.468 9 1.200 2 1.550 0 1.457 7 0.983 5 0.983 3 •P o 2.410 9 2.010 4 0.966 8 00 — CJ P^ 193 2 162 3 888 1 023 5 982 9 253 9 050 8 674 1 o ,070 5 ,242 4 ,022 5 ,451 9 o ,914 5 ,186 2 ,038 9 o ,915 0 o .954 1 ,019 2 .085 4 o ,943 4 o ,899 2 o ,833 2 o ,897 5 o ,751 6 o ,139 1 ,197 6 o '" '"" •^ '" ^~ ^~ ^~ ^ ^~ ^'" '" •^ '•" •^ •^ ro '^ Vi 00 3 519 7 795 9 381 9 965 6 646 0 635 6 648 0 283 5 519 6 680 6 014 1 ,606 3 ,555 7 ,335 8 .899 0 ,342 2 .796 8 .031 0 .757 1 .743 4 .520 6 .597 1 ,381 4 ,596 0 ,541 8 ,790 7 J O eg eg CM CM CM CM CO CM eg CO ''" eg CM CM eg eg CM eg •H O T3 00 CO 0) « 3 129 7 377 6 066 0 .011 8 .007 4 .074 9 .892 6 3 ,600 6 ,470 8 .483 0 .396 8 .449 9 .575 4 .871 7 .517 4 .480 2 .054 2 .977 2 .142 6 .779 0 .017 1 .818 7 .306 7 .567 3 .592 1 ,463 6 MJ - eg o o o o •r- P '" •^ ^ '~ '" ^' '" ''" ^~ ^ ^ ^ '" ^~ '-" '*" ^" ^" •^ ^ C o U 735 7 585 0 829 3 970 0 712 1 426 2 600 3 572 1 680 2 212 9 < ,796 6 ,637 6 ,237 9 .712 5 .947 6 .783 5 .823 6 .078 1 .960 5 .069 4 .720 3 .747 0 .872 7 .855 6 .174 8 .654 2 JJ eg eg CM eg CM eg CO CM eg eg CM CM CM CM CM CM eg CM CM •i S2 118 H < CJ •o ti O P P a> ble 3 ro -o ro MJ T3 ro T) MJ ro MJ J3 '^ 4= HJ j= J= JG ^ MJ ir^ MJ T M^ ^ ^ o M^ u •4-> a> 0 u '•-t a> 0 Um CJ 0 0 4> 0 4> 0 Leg CJ 0 a> 0 Leg u TS CU Q MH •o Q CJ OH ti e- n Vi u o cu o 3 CO ti o >% Vi r) 2 u o cu CO ti O CO M ti e- o 3 >. rus ium 3 CO ium -r^'i-^T-^cdT-^ r>-|^r<>h>a>cMincMCMcooooooo sr eo -r- to 00 ^ _^ -CMCOI^"*CJ>000000 4) >% ?iCS95"^ CO cDocoQOoa>oocDa>inoooooo ^ ^ CM o> ^r^-inoh-cD-'-oinin o o o o o o •^ CO o> CD 00 3 mincM^cocooococMi^egr^ h- r*. m •«- a> -o t-OT-in^cocoinocoo>^ •- T- r^ r^ oo !>«- o •»- MJ O COCMCMCMCMCMCMCO^CO CO T- CJ> ^ CM' d d d CM "^ "^ cd "^' cu cocomcMco'cfcocMr^ooooooo T- T- 00 00 t-coo)incDininint^ooooooo ^ "^ o o 00 CO o^ooomT-eooomcOT-o^'^co ^ O CO CO <0 3 inoococoin' oo^T-inoT-i>*.ocoooooooo h- r^ o CO C7>a>in^cocoh>a>inc3>oooooo CM CM 00 in or^r«-.CMCMincM^coi^"^CMc»coin-«- CO CO CM in inoO"*-«-cO'«-inr>-^oO't-CMinr^in CO CO o o Le g hiu m r^ d d d •*:S? ooS "^ MH oo^^^inr^f-CRco^ o o o o o ^ CD CJ> CO cDinoocMcoomohoO o o o o o O CD ^ CO 00 — incOr-OOOvOO^OOO in fx- o> CO CD CM oo CM •* « >% O00h-0000<3>OCOCOCM CD CO CO O T- O CM O O H-l tj T^dddddi-^t-^T-'T^ T^ d d d •Q ti ro "^ ^CO-t-h-OOOOOT-CMCO o o o o o m T- CO CM CO C3>i^cooocot-incococ» o o o o o m o "* 1^ 00 3 oococomT-oh-Oh-h- O CM CM CM O ^ ^ in o « -o inc3>coc3>oegcMC3>ini^ CD in r^ h*- r>. CO 1^ o ^ MJ O CM cd CM CM CM IH CU CMT^T-^T-^CMCMCMCMcdeg CM d d d ^ 2 ^ cu ^cocommh-co^oco o o o o o 00 CO CD CD f-ocMC7>i^incocMooin o o o o o h- m m o •o 00 CO cDCJ>c»CDincooegin-t- ^ CO CO ^ CO •- CO CM m a> 0) 3 inCRCRCM-T-CMCMini^CO CM C3> in 1^ CO CO CO o o T- O O T- • • • • tin u •O ti p <«- o o o o ci^ U CJ CM^Oh-CJ>OT-CJ>CMin o o o o o in CD CD oo ^* o o o o o SP c« oocooocoor^co^oco 00 o o -* •«- CO ^ T- ^ cDCDO>ooo^incO'«-0) CO "* CO CM CM <• MJ 2 o>cj>ooincoco^^h-^ CD 1^ in o CM-r^-r^cMegcMCMcdcdcd CM ^ cd cd cd p CD O T- CM d d d abl e ro "^ H 121 cor>^CMr^o>i^^ocM'^^T-ooo>CMoo^o>o>coeg^inr^o>co •^^^'«-^CMCDcO'*lno^>-o>•^o^^cDCDO>ocMcO'*oooocMO COincOCOCO^OOr>-OOin-<-COI^C3>CMOC7>00^(DCMCM-<-^-<-00 ppph>.pph,,.a>ooT-T-h~T-a>ooooa>a>oor^ppppp'^ '^'*~'^Oy-ddddy--r^dr-^ddddddddddddd coinooo>ro-cDr^o 0> CO O CO CM CM^^T^CO^'«tC0CO00O>O>00 ^•r-ooegco^co"* CM O O 00 p JZ t^h-'CDcoT-ojrx-oJOoooococor^ininoocoooocMcooooco o Sb ^CM|v.cDCM^pCMpOO(3>^inoopi^in-<-OOpCMppppi^ p CJ Q> "^'^"^CMcdcdcdcdcd^^cd^cdcdcd'^'^CMCMCMcdcMCMCMCM H MJ P^T-|^T-h-COCOCDOOOinincOCOCDCMOOCOCDh-^00>OOCM 00 •o h-wco-^-^mininmcRcDcoinincooT-c^incOT-^ooT-egcj) 0) CMCM^OOr^CM"*COCMT-CMCMr»-CDCO'*l»*COCOOOOCMCO"«-"*CO C/5 o p'^'^ppcMph^r^ocMa>CM^ooooi^coppppppp cu •r-T-T^ddyr^T-^ddT^'r^d-r^-r^-r^dy-dddddd'r^dd ro ina>CMr>-ooooQOOinoooor^r^r^ocor^T-oocoinc3>ooooc3>c3> CMo>r^cDcoa>cor^coa>oO'T-incMcoT-a>cocDina>'^cooo'^h- 00 13 C0in00T-C0O00CMinOOC0in00'^OCMr^CM-r--r-C0C3>00'^'<- c/D CU •r-pT-r>.ppoqph-cMppcMCMC7>c7>T-ooor<>:ppppr^r^ -r-f^r-^dddddd-r^'r^d'r^T-^dd'r^T-^dddddddd CM ooo>r^a>oococooc}>^CMcoo>-^cooincMCMooTfC3>c3>incoc3> •-C3>^T-cD<3>CM^cDoooinooococDr««.cor^ooT-T-inooco-<- cocj>co^cMO^CM^^r»-^CMc»cj>inoo"*or^cOT-oooT-co ^. f^. ^. '^. '^, f^-^^f^ot^^Oic cDr^^-«-'<-oocM'*oo"*oocj>ooo"*inr^c3>cDcoc3>eo^oocooo '5rcoinina>a>inh>a>coh~cDCDooooc}>coh>ocMaoh--oor<-co^ ooT-coT-oooo^T-cominr-ooT-o^r^cooor^inoT-oocoo COOOCMO^CMOT-^^00'^CMOOr^COO>OOCOCOCOT-T-CD^CM eg eg CM CM ^CMincoooooo>r*-CMCMCMinooo-«-coh*oh-CMor^r^cor*- 00 — cMh~'»-^cocDinoo>egor>«.CMor^ooc»cooh-cMO>o>incoin 4) >% ^inooco<3>i^cooO'^oO'T-i^ooooincocMO>-^incDinoocooi^ H-l tj pppp^ppCM-r-^pppppCM-r^^ppppoopp^r^ ^-T-r-dr-ddT-T-T-T-y-^T^-r^T^T^T^T^T^T-^dddd-r^d - Q coinincoocococ3>cDC3>cDcocooa>CMinoin O CO -r- O r- 00 CO CO COOCO^C3>inOI^^-<-COCDCMC3>CDCOT-000 00 3 CMh-COCJ>COCOCMCMCJ>0OCJ)-«-inT-cOh~CM^rl^~ \-J V/* (k/ *k/ VM *N \Jt f*J *J* ^^ **/ T— VJ l"» \N ^^ 1^^ ^in^ ^CO^ oT~> hr^- r«CD~ lITJ- C•^M 4) X) ^ ^ T- CM CD in CM J O cMCMT-pi^pr>.ppr«-p^T--'^pppcMCMr*-pooocoo> cd"^cdcMCMCMCMCMCMiriincdiriiri'«t"^'"^'*"cdcdcMCMCMcdcdcM ^ 2 CU C0C0^O>'«-CMCOCOO>0OinCDO>C0^I^O>"^O>r^^C0O>CDCDCM c3>inoocr)CJ>inooincococoinh-i»«-"^i^ooooincMin'*ocD T3 00 CO h-ocM'*h-oooinocO'«cfooi^^coh-^i^incj>ooo>ooocj>o h-OT-iOCMinCMCMr^inCMCMOOCDOC»OOCOCOCO"^OOOCD MJ g- T-^T-'T-^T-^^T-^-r^^^CMCM-r^CMCMT-T-:CM^T-^-r^T-^^T-^^CM^ 1 ^ ti o .J=t CJ U cMO^T-incj)"^"«-cj)T-cMincMoooinoh-cooocDco'^oooocj) oor^T-ocoincMCJ>inh-inocoo^T-cMr*-cocMr»-oincOf-o> o>cocDin'*CMincMcocMooo>cMr^ooocDooor>-o-«-coooinT- < J 2 pcocoococox-in-^T-pppp^cMpr^ph^inincooocjJco cdcdcMCMCMCMCMCMCM'^^CM'^cdcdcdcdcdcMCMCMCMCMCMCMCM 122 coocDOT-i»^cDT--.-inr»-cor^coooh-coc»o50coif2SiJ£JS ;-CMC0CMC3>C0C0C0'«-OC0inT-0>C0r-CDCMC0C000O;*0>fl0CM §2 coegt^o<3>ooT-coooini-oO'«-">*or^i^egcMcog)cor-inr^o pinocococor^coi^cocD^oooooooooo>pppf^ppp ^ I^H ddr-^ddddddddddd-r-'T-r-^ddddooooo ai ai o ai 1^ CO m ^^CMCOm^CO^O^-CMCMO^-O^— o^^ m h* m c»>inc3>egcococoocoococoincoor;--inco_. -. o CO CO m r^ m-, cjw>. CwCOcoi^mcomi^r^-^^CMin-^CMTrcocM;^O CO CO h- © ^— C-..-w.i^ocoincMooinT-r^oocMCMcoppppi^pi^i^O T- o p cjCMCMcdcMCMCMCMCMcdcNCMCMCMCMcdcdcdcMcdCMCMCMcdCMCMc> m 1^ o> d Telso n Lengt h coooo>oocD'«-'«-CMT-^inT-cMCDCoooo^inT-T-cDincor^in 00 "O 0)inCM'r-CMCM00r^T-CMinC0C0CDT-CMinr»-^OC0'^C3>O0)h- r^inr^oococ3>T-incococoo>CMooi^cocD^ocDr^cMoocDco pppph--h>.ppoocDin^coooooppppppppi^>^>^ dd-r^ddddddddddddddddddddddd 3r d An t r^CMcoooT-cMcoinc3>ooooincMcocot«<-h-incor«-oocor^i^co cooocoooooocOT-oooooooor^incor^ooinoinincMOi^oo cococj>coegooincMCMOco^co^^in-*r^r^cMcooinoocMr-- Pe d ppf-pr<>.h>.|v.r>-ppp<^Qor^ooooc7>ph«:i^r^ppr^pp Se g dd-r^ddddddddddddddddddddddd c^ CO a> o o 00 h>- i^inoo"*"^coc»inc»cooocM^^r^cD^'^in^ CJ m CO CD ^ 1^ ^ CMin"*CDC3)O>CM00T-CMC3>inCMC3>CO00inC3>'«-CD % ti r«>incMinoocM ^ooooc7>ooooooooino>cMOO<^^c3>r^oc3>r^ pT-pcM^incMca-^^CRoococoinin"" ~ ~ ^ CMCOT-CMCMCM-»-0OCM T-^ S p/ D x> Umo ruo CU ^_, C3>CMCOOO^COr^r<~OOCDCMCO^OOOOOOCOCM(3>a>CMOOOCMOI^ cocDinr^^cj>r-in^c3»cDinoo"^r^c3)^-»-cooocMt^r>».c3)ooo 'P M OOCD'^CM-<-C3>incOOinCDOOCDI^COCDC7>OOCD^-^OOCOOCDO ti 3 r^co-«-co^"^incMCDcOT-o>^h-r^cD'«j-co"^cocO'^cj>CMCMoo :^ •r^T-^CMT-^'r^-^'T-^'r^T-^'r^-.-^dT-^T-^'r^'r^'r^-r^T-^-r^-f-.r^'r-^'r^'r^-r^ ar p Ul o ro •* m OO o CO m h- CO o 00 CD oo CO a> "* CD CO CM o Oi ^ 00 at 1^ CO CD ^ o CO CM ^ 00 CO o "* "* h- o o m in <3> CO 00 ^ ai 1^ in CO o -* m 00 in m -* CM CM o T~ 00 o> Tf C3> 00 h- CO r^ m CD in "* CD a> >^ a> 00 CM o> p •r^ a> ai T- p ai 00 00 p t-^ eg p p p r^ T— oo a> h- h^ MJ t> T— '" '" ^ ^ ^ '" '" ^~ '" '" •^ JS ti d d d d d d d d d d d d d '" ir> Q m 00 r^ m r^ •* r^ ai ai 00 CO o o o 00 1^ CM CO 1^ a> CO o> CO CD CO CO OO y— in at T— ?) CO 00 T— CM ^ 1^ 00 C3> CM 00 ai 00 3 0o0 eg -* in Oi ^ ^o h- r^ h- CM ai in 00 -o* ^ in CO 4> T3 CO ion "o* CO h«- T— ai ^ aieg •o* •r-^ CM CO m CO at o> aieg HJ O o eg m o o> a> p p o m CU cd cd "^ cd cd cd cd cd cd cd CM eg cd cd cd cd -* eg cd eg CM cd cd eg cd cd O m Ul OH CO 00 CM CD •* o CM •* CD oo 00 -^ 00 o CD o oo CM CO o CM -^ o CM CD ai •* CO "* h~ 00 CD ai CO CM in 00 ^— h- in in 00 ai •o 00 CO eg CoD m CoO 00 m a> CM ^ 00 in ho^ o in "^ CO CO CO 1^ ai o ^ ai eg o o o> eg o r^ o> o a> a> p p p •^ p p h- •^ -^ eg CM CM 1^ p CM "^ CM p p h^ p p p 2. 0 g J & 2. 1 2. 0 ^ ti o u CJ in m 00 h~ CO ai 00 a> co in CM CD ai eg CO eg ai CO T— eg l>. in in CM in T-H 00 CO o ai CM a> 00 -* CD "* CO CD CO -* r^ CO m CM h- CO 00 h* 00 CD r^ C3> CO ^ o a> CO "* •* 00 CM o "^ o ^ CO oo h- CD 00 CM T— <^ co 00 r^ iJon CO CO < CM "* CD a> •"d- o 00 in CO in CD T— in CD oo ^— m -* a> CO "* -* -^ m "* o M^ 2 CM CM CM CM CM CM CM CM 2CJ cd cd cd cd eg cd CM eg CM eg cd cd eg eg cd eg CM cd H abl e 123 o -«»• -^ CM ><- r^ cocMoocj)^cgincocor^inor- oocDcoo^coO'»-CMcor^r^ocoo>r^cocM^o>CMcoO'^cMoo oo^OT-i^^o^coh^ino>oj'^T-ino>^ino>coincD'»-pp P .p C0C00)^r*»^CD'«-T-CM^I^OCMCDC0i0CD^0>0)CM^h^'^C0 o ^ ppp^o^cj>h-inincDcococMcO'«-o>T-T-oh^'<-;Pp^p CO p CJ a> CMcdCMcdcdcdcdCMCMCMCMCMegcdcd-^T-^CMCMCM-r^CMcdT-CMCM H M4 h*in^T-oor«*incj>o>CMcoin-*-CMinT-cooooco^cMCMCDi;^ 00 •o COCOOCMCOCMinc- - -' O -ocj>ocMCJ>^CMCDCMegoo^cj>r^cDOT-co a> 4> ^ooTfo>ini^coo>cO'^oincOT-ooocj>ocMC3^r-r^co'»- C/) CU ph-;h^pph-inoooo>r^T-^oooo)r^pppppi^^pp ddddddT-^ddd^dd^d''*^'"*'^'^''^'"*'"''^'''^ ro ^h«.C3>"*000<3)T-r^'^in^COT-CMOOCDCM"«-COOOO-<-COCD« coT-coO"^-f-co^oooo>cj)oor^oo>'^cococor^CMooco<3)^ 00 13 ooinoot^C3>ooT-^cD'«-co^ooo"*'*r>-r*-cD^cor*- •Uol «-pi ro OH • m^ coinf-cOT-ocM-»-''*cOT-i^T-cocMooooincM 1^ CM C3> 00 C3> oo T-OOT-O^mOOOCROCDOCOCD^r^OOO I^T-C»OI^CO^O ^ 22 oooooco<3>h-CMcoinooinin egcor--egcj>cMincMr«-h--«-cocj" ~ > c^ 3 cocoinin^cO'«-inT-egcoco O ^ CO T- •*- "o CMOOC3>COOOOC3> CM •^' t-^ T^ CM •^H o^ ro egcDcocMinoin^cj>ooco'*'*ooco^cj>c»CMegincDoooT-c» 00 ^ CM^OCOCMCDr>-CMinOCOOOCM^inCM^OC»CDT-CMOh-T- 4) >^ CM^ooot^^coh--«-r*^t^o^o^'«-cD<3>'*^cj>T-^ooinT- ^ tj cj)cooegh-;"^r^pppp'r-p-r-cMppppp^pi>».^inin •P .^ d T-^ T-' •«-: d T-^ ^ d d d -r-" T^ d T-' T-^ T-" d d d d d d d d d d ir> •^ T-0000CM'«-'«fT-0>00CJ>C0^C0CDI^^"«-inCM00OT-OT-OI^ a> cMin^CM"*cocoooinoinoc3>cDinT-c»^r*-OT-oocMoocM 00 3 cooegcDCMincococomo^ooocMcoino^r^ooco^^cDO (U -o J O COOOCOppCMCM'^CMp-r-;ppppCMpph-;PpCMCMr^CJ>00 _C CU cdcdcdcdcd"*'iocdcdcdcdcMCMcd'*'^cdcdcdcdcMcd"^CMCMCM CU ococOT-ooocMOJOoegr^ooooocoint^inr^T-cMoooococo cDincooocoinr»-cDinT-^CMoooocDh-cDino>^CMT-o>ino>'«- •o 00 CO cocoooooooco-^^ooegoocRr^^h-cOT-i^ooooooT-coeg 0) 4) 3 OOOCOint^OOOOC3>h«;ppp'r^pi^CMp'r-ppp^h..oOO>r^ J J & T-^CMT-^i-^T-^T-^CM^-r-^^-r^^T-^^CMCMT-^CMCMT-^T-^CMCM^T-^-r-' ,^^ .P ti C^ CJ O CJ •r-o>cocoT-ocor*-ininr^cocMT-o^-cMcoocDcoo>^o>cDCM coinoococj>cj>cocMincooocoocDcoinh-Oh«-ooT-cocot^C3>CM ^ "*h»ocj>cDCMr^egcMCJ)-<-inoooo"^CMCDCMh-CMCMincoooo^ SP »3 pT-pococo'ttegpppppT-^pr^^i^egppppT-p^ < CMcdcMcdcMcdiricdcMCMcdegcMcdcdcdcMCMcdcMCMCMcdcMCMCM JP abl e H 124 '" 1^ o> T- ^— m ^ cocDOooinTrinh-cDoo'^ooco^;j£4CMO>in CO ^ CO 1^ r^ o> ^ "^cococMcooo'^'^'^r-eginjCMCMoowcoco '- CO in CO CO 00 ocDCMi^ocM'^oooT^r^po^oor^r^^- egpcMpeocMf^cocoincoinr^cocDCDincDppppppP'^ '^T^^ddd dddd dddd doooo ^ioor^in'«-cM^ooc3)'^coinr^ 5fCM5CJ>coinin^cDooo m CMinmr^ooT-t^i^o^ocooo p ^ r^inininoooino>oor^cMr>-co Si!5^S55 cop^cj)^^^oc»cDooc3>c»^coinT-cocMCMcOT-r^h-T-op "*h-0"^inooininin^cMCMO)h-ooocMcocMOoo^cj>cDcocM 00 T3 egcoo>ooin^coin^coo^r^^^o>T-coi^cDCMcoooooT- V 4> pxhincoinin^cDincocoooh^r^r^r^cDt^^ooa>T-|pppp cn O. ddddddddddddddddddddd-r-y-ddd ro "*cDr^coocj>r^oooocDoocM'^co^^incMr^ocDOT-cMin^ ininocMT-cOT-inoooocM"*'«-incMOCJ>oooocoin-«-o>o>o>'«- 00 13 r^^cO'*^r^CMO>coocoinocof^cD'*-in^in'*ooocoooo a> 4> ^pppppi^'^t^mcor^t^i^ooh-coppppp'^. ppf*: ddddddddddddddddddddd-r^-r^ddd CN in eg in CM C3> C7> ^•*CMr^cMC»cDinr^oooincoc3>cO"*co^cDO CO "^ 00 CD CM h- inOinCMCJ>OC»OOCOCOCMC3)COOCMr^CMOC3> X ti CDC0I^CMCD-r-00O"*C0'*O-r-CJ>C3)OC0r*.CMCMinOl^inr^C0 coincMcocM-«-r^cD •* CO CJ> CM T- O <3> O O ^O^"*CDC0CMinCM y^ T^ d -r^ -r^ 1-^ d •o o ro OH T-incDCDcooJcO'T-ini^coco^T-co^oooo'^oocDcooo^'*- ooooocoinco^^"^ocoooT-^cocoor^i^coh-i^ocoh«-r«. 'O Vi CM'«-^inoi»-"«-OT-r^oc3>oocOT-flocM'«-incD^CM"*inoin cDr^r^inooc3>cMr-cDoo^'«-cocO"^"*C3>CMco^r*.OT-inh~CM eg CM CM •H CJ ro ocoococooo"^'*T-^r^oocO"*cocO'*-c3>cO'^T-T-ooincMco CMh-T-h-cocMCMini^T-^in'^ooco-'-ino'^T-cocoincO'^co 00 —. cDcO'«-coinc»ooinocj>r^oococMCMoocD'«-coincMoocoT-oocM 4) >^ eg"«-coo^o^oo^i^t^o>i>^oc»cooi^cMco^incoO"i- y-^y-^T^y^^^y^y-^y^y-^ddddT^dd^dy-^y-^^^y-^T^y-^ IO "^ CM^cocj>r^CMr«-coooooincj>ocMcoin-«-cMC3>cO'«-^cocj>r*- co^^incj>-<-"«-r>*coi^ 00 3 T-inoom^cocOT-cM •^ocoocMO-r-^^oc^T-incoooT- 4> •O ppcMppppp'd; OOCJ)CDh-^^CM^CMT-OOOCO^CMOO MJ O cocococococo^coco ^cj>r«~r-r«-coincDh«-T-oincMooc»T- •^ 2 egegcMCMT^CMegegegcd'^'^'^rcdegcd OH egcMCMr*-c3>cj>CMoofloinor^in^cooor»-r^oocooocMincM'r-cD OCOCOOCMC3>CMI^OO^"«-incOCOT-OCOincOCDCO^CMCDCOCO t3 00 Vi 4> 0> 3 •r-cocM^CM^'^oococMr*oo-^cO'*cocM'*co^inT-inh-incj> CDOOOflOOOOCMr^h^p^CMp-r^pCMp^pppCMT-h-cOin -r^eg-t-'^T-^CMCM^^^T-^^^^d-^-^^T^^T^^CMCMT^-r^T-^ J ^ ti P .:^ CJ O CJ r>*-^o^"^cocj)inr«-T-ococMO^"^^oocoin"^"^CMT-cj>in co'^cooococooocoinocMCM'^egcor^r^coeginocDcO'^eg oocMCMincO'<4-inoooo^c7>cocDoooo>CM^oa>0'^cocor^-r- < h-pr>«-oo^ocDoocj>CMcoi^p^ppr-;CMt-;pincor^coi^p egcMCMCMcdcdcdcMCMcd^cMCMegcMCM'r-'cMCMCMcdcdcdcdcMCM 125 OOCMr«-OOCO'»-^COO>^T-COh-^CMOOCD'^'«-^'-'<~COp p ooinincDoocMCMT-r^ocMcocM'SrcMOoococoococoincM^cM o S ^S2Ii2r:'*oo>ino>«j-cj>'^incD'*"*oo'*"^cj>inT-r;-incM CO ppppppopoo'T-inr^h>coooa>ooo>pi^ppp'^p X oooddd-r-dr-T-^T-^dddddddddd^-'-''-''-''- 4> s moocMocoooeginT-co^eginr^cooor^'^OJCoincoi^ooo r"2^T?^«0"*inococ3>cocDO>r^incM^ ^cocoinooT-oooooT-o>r^t^T-cMor^cMoo^oor^ino>CMO ^ocMCMincDegcDoo^r^o>CMcor*-'«-r^oooincococoooflO 00 13 CD00"^00T-00"^"^OI^C3>C000r-T-OOr^T-CMC0C0'<-egc0O C/!) OH ppppppcMpcMocMooo>ooooppegpr-:pT-;p^p 000T-00T-0T-'*-T-00 -r^d^t-^d-r-^dd-r^-r^^"*-^ ro 00 "O 4) Cy!) OH CN I^CMT-CMT-OC0C0CJ>T-00CMC000r>-T-in0000inCMC0C0CM'*l^ -a CJ OOCOCOCOOOr>-r-CMOCOCOCOCMCMOI>»-inOOh*CMCMCMCMOC» ti CMCMincOOOCOinT-CDT-COCMI^OI^CDCMCMCMT-CO^COCDincO o--> O - CJ> 1^ T- ocMCDooco-^^cMCMT-'t-r^cM'^CMOinininpp :^ p/ D •Uol oUl ro CU ,^^ C3>OOCOOOOCJ>CMC3>T-CMCJ>^"*T-COI>^OCMCOI>^COCOinCD^^ incj>ooooc3>oo^r^cDOcooococ3>h*coT-cDO)h*"^ini^co^ •>? CO oor^f-cMCMoocoinc3>coooc»oocM^CMr>-^oo"*eoh>-cj><«-T-co <^ 3 •»-cMCMincMcoin^o>o>T-cMincoincooocoincoco^cDininco ^ T-:t-:-r^CM>r-:"r^-r^-r^-r-:T-:CM-r^i^T-'-r^T^-r-:-r^'^-r^-^'r--r^T^'r^T-: ar p t3 ro u ooc3>oo"^^cocOT-inininT-i^h-cocooor^incj>oooocoor«-T- l^inCMOOh*.inCM^CDCDCMCM^OCDOCJ)inoOOT-OOOCMT- 00 c3>^i^coT-cocoT-r^inoin^in^incocor».r«-cMco^cM^co 4> >. h-C»C3>COOCJ>CMOO^CMI^^<3>CMC3>Oin^T-C»in^inT-CM'«- MJ t3 O O O T- f- o OT-T-I-T-O'f-Ot-T-I- O O t- T- T- -t- f- ti T-coi^i^cj)cooh-CMco^cocoin'^T-"^cooi^r^o^ooT-oo C» incocj>"^t*-r^-«-cDi^i^ooincocooooooococMcoin^"^C3>^ 00 3 0>CMCOCMT-COinCMT-COOh*CJ>-<-r»-^OOC3)COO>OOOOOOCDCDin 4> ^3 inCD'^pi^PppCMT-;pp"^pppT-ph-.p^pp'^CM"^ MJ O CU CMCMCMh^CMCM'^CM'^'^'^CMcdcdcMcd-^cdcdcdcdcd^cdcdcd JZ O Ul m CU oocoocDOT-inoO'»-^ocoincMCJ>oooocor*-o>o^inoT- T3 00 CO Oh-h-^CM'T-COOOT-r*-OC3>h-OCMh*h-r-CMOOOC»COC»^ 4) 4> 3 cooT-in^T-T-^o>^'r-^-0)T-'r-o>OT-i^coinoinh»-o>oo M4 egcMCMCD^CMincocM<3>pi^ppppcMh>.pppiv.coin"^"^ g B- ^^T-^CM^T-^^T-CM^CMT-^-r^T-^-r-^^CM-r^CM-r^-r^T-'T-^-r^T-^T-^ x: ti O u CJ .•CMOi-CO^CMinT-CMCMC^T-CMT-OOT-COh-T-^OO^inh^O) ^' coor^ininococor^^oocom-i-cDoooocoi^ocDCMO^CM"^ 00 Vi ooinocMoooooinr»-o>cj)oo^i^o>co^^ocoi^cooot^i^ < MJ 2 CMCOT-opcOT-^inCMppppppppppp-'-;pCMOT- 4> cMCMCMdegcMcdcMcdcdcdcMCMcdcMCMcdcMcdcMCMcdcdcdcdcd ir> abl e H 126 ^ r^ "* eg ^ inh-in-«-cocj>c3)00'^0"*cor»-5oocO"* ^T-coor^cocDooooo>r«-oocMcoino>'*inr^h-h-<0'*o>r;^ CMcoinr^cocMT-ooor^T-r^r^for^^incoO'*-cMCDcO'^T-in p JP oowin-^cD^r^cocoegoocMOi-CMcocoinco^^gcDh-T-^ o 00 ppcDT-0'«-ininincooT-T-'«-ooT-cMT-cMp^ppi^pp CO cd^^^^CMCMcdcMCMCMCMCMcdcMCMCMCMCMCMCMcdin^in^ 4> 4c> H HJ CMcDoo^oo^mcoooo^^cocMco^ooooooi^ocooin-*- •«-oocMcoinoococoin"*cMC3>^T-oooc3>cj>CM^incoinT- 00 "O cocDr*-cDooo>cj>cM'*co^r^r^r^^c3>r^cMCMr^coocor^5to a> 4> pT-'r^pp^r^ooph-in^^r-^h-;p^ppppi^ppP'7 in CU -r^-r^-r^dy^ddddddddddddddy-ddy-y-dy- ro I^COinT-CDCOOOCMOCOOOCDCO^t^CD^COOCMOOO>inT-r^ oa>a>cDCMa>'^ocDa>oooocoincDO>CMr^cMooo>^^oor^^ 00 13 0) 4> CMCMcOt^CDinc0^inT-OOinCJ)CMC»0'»-CO^COCJ>f^CDCMOO CA) CU pcMT-^T-ppr«-.pr^ppin^r^cO"^pppegpppcMpp •r^y-'T^'r^dddy^ddddddddddd'r^ddy^y^dd CN r*-cocoo>'«-incor«-CMO>^o>cDincDoor»-T-r«-in^cocOT-^'^ COOOOOCMO>r^OOCDCMCOOOOOCMO^h-inOCOCJ>T-CDOC3>COI^ r^r^coinoococoin'*cDinoooocMT-inr^cococMC3>cooocMOCM ^coooinooT-inoT-r^ooh-cocMoooooo-^oocMOcooop T-^i-^cMCMT-^d^T^T^-r^dddt-'^ddi-^ddd^'^CMT-^-r^ r^ooin'^ooooooo^C3>CMinoor^T-inT-cDooocooo^i^T-co inincMr^coooa>cDinocoh>cocM^co^incM^cococ7>cocMin cDro-CMoor^oocooininr^CMooooc7>inT-coro-h>-cocoooi^cj>o ^oooc»inr^oc3>c3>T-inoor^coT-oocj>cj>oocDCJ>coo^h*o ^^CMT--.-OT-T-OT-000-«--«-OOOOOOr-CMCM CM CMoocMoocMoocor^cocj>CMCJ>CMcooT-T-oegcor-inr*.egc3)"^ 00 » incOCMCOT-COCM^CM^CJ>r^'*CJ>0"*CDCD'*CJ>C3>00'*OOCJ>l^ 4) >. cot-ooo^ino>ocO'«-r^t-r^ooin^^r^incO'*-T-cocoooo J tj or^r^h^in(3>ocDoa>cDocooooor<»r^c3>a>oooinh>^in ^T-^^'^T-^d^-r^^ddT-^d-r^^dddddd^-r^T-^T^-r-" t^CM0)0>^'^-h»'0>00C0000)O'*C0OT-'0>^CO'^000>T— ^o> T-cO'*"«-oh«-i*-oor>*cocor^coinoocMinot^or*«-- '•^ x: ti G 1?^ CJ O CJ oinegoocOT-cj>T-incoinr^inocO"*o>or^ocooocj>"^co cMcDooc3>co'*cDOcoeg^cocor>*CMO>coi^coT-T-f-^oo^oo OC0I^^C0'*Or-^CMr^OCJ>^'*h-C000CDO^C0CJ)C3JC3)O < pp'^T-coi^ocMOOCMh-pp^ppppi^por^r^incM JJ CMcd^^cd-r^egcd^'cM-f-^-r^T-^CMCM^-r^'r^T-^egT^eg^inTf^ 127 0>OT-T-CMOCMCMC3>CMOOOOOO r^ ^ 00 eg CM00T-ioo^CMCJ>in-.-OOOOOO CD O 0> 0> CMr>.or^CMcooocO'^oocDcoincDin oo CD ^ CO o <^ pph>-ppooo>ocMT-oorooooopr^ 00 CM p p ja X dddddddy^y-^y-^dddddd dddd ^ •«H cMr^cooococomi^ocooooooo CO CD CJ> CO cDinh-co^coi^coincooooooo p JG CO o> r- '^ ^oocMcor^oocDcomooooooo CO CO 1^ o> 00 TJ r^-«-cocoooc»cMr^inc3>oooooo CM oo O) 00 U CJ oocMoor^cDa>cocMocoa>cDcoT-oo> ^ m 1- CO m cu cpppppppppT-;coppppp 00 CM p p dddddddd OT-OOOT-OO dddd ro CJ>C3)CM"*OOt^"*-'*CMCOOOOOOO 00 00 ^ CO 00 "O h-oot^^T-oooocoh-cooooooo C3> r^ r^ ^ « 4) in^T-cj>incM^cD0O"*coco^cOT-cM O CM T- CO C/!} CU ppppmcor^h-i^ococo^^-*-^ 00 CM O O ddddddddddddy-^y^T^T^ dddd CM lOCOCOCMCOCOa>OOCOCMOOOOOO a> o> oo o CJ C0CMC0CMCMCDCDin"^C»^OOOOO in m ^ o> ti co-«-incoooooincM"*cj>"^o>oor^r^cM CM CM CM ^ 1ti p-r^p-^^p^pppT-cMcomcoh- CO CO o o •r-^T-^^'r^'r-^-r-^T--r^T-^^d-r-^T-^^T-^-r^ T^ d d d p/ D •o UM oUl ro CU , oococj>r^ooooi^^cj>ooooooo o CO 00 1^ cocor^cMinincMOC3>c7>oooooo t^ o ai 00 CO 'S C3)0OCMC»COini^l^C0CMCMCM^CO^^ C3> C7> CM in ti 3 ^ CO o o int-ococMco-«-r-C3>i^ooincoor^o ^ y^ d d d ar p T3 CJ ruo <3>CDinCM00Oh-"*CJ>"^O o o o Oi eg CO a> oa>cDr<>r^o>r^coino>o o 00 CO CO m 00 cDincoincMC7>oocooocMr» o o CO 00 o CM -* 4> ">. CMoococj>^oor-^cocor»- rC^O CoO CO o CO o o MJ ts "T^ddd-r-^dd-r^T-'^d •«-•<- T- T- o o o JS ti •4J inocoococoinininT-oooooo O O •«- CO CO oa>a>cj>h-cocM^C3>ooooooo h- C3> CO ^ 00 3 cDOCJ)ooin"*cj)T-cj>^in"^coincoco CO in t^ "* 4> ts CM^t-h-CD0Opppr-;^^ppCMp "^ p p •«- M^ O cdcMCMCMCMT^cMcd-^cd^^iri'^'inin CO d d d ^ CU •4-* •r> oUl CU r<-coooT-^o^r^cDO)oooooo 1^ o CO r^ m CO m o> T3 00 CO T-CMCDCDT-h-OOCMT-CMOOOOOO T- CD CO CD coo>ino>r^cocMOo>coo^inincM^ r^ •* o o a3> a> 3 h-^CMcocoinmoooooooocooocomco P J & ^'T-^-r^^T-^T-^T^v^r-^T^CMCMCMCMCMCM y^ d d d ^^ JZ ti P O u U cj>inr«-coocooT-c3>cooooooo CM O CD CD CD T- CJ> h". —; 00 V) OC3>OOCO^CDO>in-<-000000 m oo m T- cMcoT-c3>ooocMinegcoo>r^incj>c3>oo C3) h* O T- <: MJ 2 ppppcoc7>cDpp^pcMh«;pegcM CM d d d 4> cdcMCMCMCMCMCMcdcdcdcdcd'^cd"^^ abl e H 128 ^oocoot-moin^ooooJco^oo^^ooinooco^OT- 0"^^cj>h--«-r«-'^inr^oooT-T-cMooc»coT-i>-^ocDCM^ T-r*>r^r<.cocDcoooh-in^oocDcocoa>cocD'^CMCDoocoinoco C u- egT-coincDoocoooinoor^cocDcocDCMCMcop^cop^^r«-in JS X y^y^y^ddddddddddddddddddddddd o>cocoooinT-'»-cDh*-^ooo>oooininco^r»-o>cM'r-T-oo C0r*-C»^'r-i-OinCD^OI^C3>'«-CMCM^"*CMC0'«-^C0OT-T- CMC3>O^T-C3>COCOCMOOr-r»-COCOCM'*0>r^h*OOCMf-OCO^CD ooco^oooocDoop^CMCMppppph^r-;Ppr>;ph-.h*-p CMCMCM-r^-r-^T-T-^T-^T-^cdcdCMCMCMCMCMcdCMT-'-r-^T-^-r-^T-^-r^T-^-,-^ OQ ^r>-<3>CMin^h-coo>'«-inO'«-ocMh-co^^r-r^in^inT-oo oo^T-T-^CMOh»-^ino>CMcoh-CM'<-in^"*ooinc3)r^oinoo inocMT-ininr>.in'«-^coooot*-ococor^cMooo)cD'«-o>oo c MJ cj>OT-pr^h-.h-ppppppT-h-.pcMpp^pp'*ppp 3 CO X P O y- y'dddddddy^ddy^ddy^y^ddddddd d 4) yn O H 2 cj>v-oooT--<-cMCDoo^incj>r*»oocO'^incoh>-in-«-h*h-oor^co inCMOCMC0I^O^C0C0h~r-O'«-C000"*T-CMC0CJ>'«-C3>C0C0^ O ^ cDininincoc}>ooincococoooinegh<-ocMoo^i^i^cMoocOT-cM < J2 X ppc3>ini^r»ini^coc7>ooina>h>coincoi^pin'^p-^cDcoco 4> 4) , dy^dddddddddddddddddddddddd 129 OOCOCMinCDCJ>C»COC3>r»-"^CD'»-OOCO"*'»-r^COOCDT-T-CJ>"*CM i^cDCj)CJ)CJ>ocM^oO'«-cj>^coinT-r*-ooocM-*-ini^oo^i^co O H •r-OCOCM^CDCOCJ>C3>CMinO>COI^"«-COincOC3>COCJ>0'*r*-00^ ja X h-.inT-CMCM"«-CMT-p^f-;^-r-CM^pppppCMpp'«-;i-;CM 4) , dddddddddddddddddddddddddd CMOCMOinCJ>inC3>O-«-CM00CJ>CM-«--«-inc0^CJ>inC0-«-00CDCM cDCJ>CMr»-cj>T-coooincoocMCMi^cocor»-cMO>"*inoooooincD h-OOCD^COOCDinCO^I^OOOOC»^COO>inoOC3>T--«-OCMOO^ pini^oocor^csj^^ooincjioo-r^cM-r^i^pppcRinT-oo^ •r-'T-'cMCMCMCM'r-^T-^CMCM^T-^T-^T-^CMCMCMT-^CMi-^T-^-i-^CMCM'r^eg OQ ^o>ooooin^oo^O'«-egcMOcocOT-cocoo)0)CJ)CMcoco^o •«-inT-ooocj)ooooino>r>-T-cMh*cocDT--«-h*-incM'«-inco cDCDinoocM^o>o>cDCM^oooo>CMcor^incD^cMO>oeg^'r- a> ^ MJ 3 in^oocDinpppppppppi*-;pr^, ^ppp^^-;'*p'* P Vi X dddddddddddddddddddddddddd « "^H O H 2 CJ C7>C3>CDCOCOOOOininoOCMO>CJ>OOOCDOCMCDC3>OOOT-CM cDO'*egooh-CMcocD^'^r*-cocDcoincOT--<-cj>cocMooooinr^ ino>T-cj>ooo^cj>"^cocoo>co^incM'«-coc3>o o eg -r- CO CO CO If) ' < O ^ r«.^t^-5fcD'^in'^in^-^coinini^h»a)CDcocococDCD^^co J2 X dddddddddddddddddddddddddd i4 > 130 r^r»-ooc3>^T-cMOO cD-«-co-t-cj>0)"*ini^coegcooooT-oco 00I^CO00t-C3>"*'^^ cocococooo"*cO"*i^cD^r^"*oooocoin C}>COr>-CMCMCDinC3>0 C0-«-^CDCJ>C0C»CDr^CMC0^<3>OCD"*00 O H •«-CMCMCMCMCOpppppp^pr*;PCMCMppCMCM"^CMCMCM ^ X dddddddddddddddddddddddddd OOOCMCMCMOC5>h«-h-T-T-cO-.-COCM"*CDC»COCOCDin^CMegO incOCO00C3>^COCMh-CJ>CMT-0OCMO>COr^CJ>^0OCM0O'«^in^CO p ocMcOT-coc3>cocoooc3>'c»-cor-^T-coin"*coincM^oocj)coc3) o S T-incMinco^T-cD^^mcocooo^cocM^cococM'^ooinco CM CM CM CM CM CM CM -r^-r^T^^T^-r^egCMT-^^T^-r^-r.^ eg el s X H n CMh-»ncDr^incMoocooooocoi>^oo^cj>co"*oor^cocj)cO"*egoo 0"*OOCOCJ>inT-CM^C3)C3>h^incO'*^h-00' 131 incMcocMT-oh^coincocMcOT-oooco '«tf-inincMCMro-CMooa>co u»TOT-r»-coinoo"*-«-o>oococj>oocOT-c3— r*«.»«*^aj-^-T— WJWjcjwJuwfi^OJ^r^-^cjJ^^OJOJC^OJt.> j r*-ocMoooocD"^^incDCJ>CMr>«-c»co^r««-T-T-c»OT-i^ocoih»T— 0>^0>0>COO>n0 O H oocMoor-cRoocMoooocM^-t-egoegcoegcMT-T-inegcocMOCM ^ X 4) T dddddddddddddddddddddddddd C0CMC0OinCMinCM"*l^00^^CDOCMinC5>O CO m CD CO CM h- 00 ^oo^oooh-h*eginh-h-r«-inoh-r>-r*-c3>co m •«-•*- CJ> C3> CD coc}>h>-^coh>-ooo>^inc7>ooo>inoocDincDoo^ininr^oooo CO »s^ pppp'^ph-p^pppppT-;|«-.ph-ppCMpCMCOT-00 •r^i-^^^T-^^T-^-r-"i-^f-^T-^^T-^^CM^^f-^-r^CMCMcdcdCMCM-«-^ CMCJ>^<3>o^coooc3)OincMT-inoo^r-CMOoocor-cocM^^ ^oooh-T-cj)CMO)ooT-cMooooin"*o>cocoT-ino^r>.ooo'^ p J h«.CMOOOOCOC3>inCMr^OOOC3)CMCOC3>CD"*OCMCO^incOOCJ)0 h>oor^inr<>;ppppppph>:ppppp^i^ppoooor^ g ^ X ddddddy^ddy^dddddddddddddy^y^d O H 2 CJ h-oor*.oooo"*cMx-inocMininoo"*oococococMr<-CMCD-«-inco c3>CM-«-oo"*oocD"*inoor«-.cDO>coooin"*cor^r^"*coh*or^in o ^ CJ>CMOcOCMcOCOOOinT-'^CJ>COCM^T-CMt^T->itOOT-C3>'^"^ ^ X P'r^po)ooo^o)o>oin^^^inin^in'*inr*.copinp^ 4> -, dy^ddy^dy^ddy^dddddddddddddddd 132 CMinT-CM^CRT-^h-OCM-r-C^COininh-CMh-CROOOOI^COini^ OOJt^T— 0>OC000C0O00'«— CJ0^-h-T-^-^C000h*000)Oin00 C5 UH CDT-CM^inCDCMCO^OCMCM^O^h-incOinCMC3)CMCJ>^^l^ CMCM'«-CO^^CMCMCMCMCMCMCMCMCMCMT-cOCM't-CMOOOOCJ>CO ^ X 4) ^ ddddddddddddddddddddddy^-r^dd ^Oin00-«-T-CMI^I^CMCMCMCMr^h«-h~h*in-«-CMh-00O"*T-O oooi-cocDego^ininT-cj>coc»coininT-cocj>CM^ocMooco coh>ocDOooa>oocMO-«-oh-^cooc7>incj>CMooocMinoin ^pph~ppCD'<-;Pp^-<-;p^p^ppCMppppppCM -S X T-^T-^-r-^cd^^CMCMCMCMcdcMCMCMCMCMcdCMCM-r^T-^-r-^CMT-^^CM inoooT-h-incocoh-cocDoor-^ininr*-cooT-inc3>T-cMcooo cMi^incoo^T-oo>cDcO"^"*"*^^T-r*.or-.cooooooooooo a>oa>r^in^oocMinooooO'r-a>o>oooo^cDcoo-^^'^oo 4> ^ MJ ^ppph<>pppppp'^pppr<.c7>i^oococDa>h-c3>a><3> CO X ddddddddy^y^dddddddddddddddd g 4) «^ O CJ coincocMcoco^T-a>in^oO'<-oocMooinr^co^^cooo^coco inr^oo'^ooinincoocoo^inin^in^-^CM'^c7>egcDi^coc7> cOT-coinco^cDinoa>cocDOOOT-^o>oincDCMC3>ooooa> < o ^ '^p^h^p^h^ppcDi^^pinincDcococoininh^oi^r^co ^ X dddddddddddddddddddddddddd 1 133 cO"^cococO"*CMT-^o^cor^"^h>-inc»ooincMT--«-inincocD T-cDcoocoegco^"^cocor».oooocDoo^cMinh*CMcocoor^ or^coinoocMCMoocMcocOT-T-ocor«-h»-CMooc»cj>^i^cDinco G UH ooc»OT-ooin^h*inininco"^in^^cococMincoinT-in^T- JS X ddy^y^ddddddddddddddddddy^y^y^y^ ^ 1 oegcj>oco-<-'^oooin^in'«*cooooooegc3>"*^i^ooincDO or^oo^in^h»ina>ego-r-c7>ina>cMcocD^ocococDC7>coin ^oi^h-0'*inr»-in-«-CMCJ>coooh*inoin"r-coegf-cD-«-ooin p 5 pppCMCMT-;"^cDCMcoooT-o>h-eg^cMincDincj>cDoococo CO S/< CMCMCMCMCM-r^T-'T-^ir^T-'T-^-r^-r^T-^-r^T-^T-^-r^-r^-r^-.-^-t-^CM'r^CMCM H PQ r»-^'«-CMinc3)CJ>inincj>CDCDO>inh^C3>r«-cDcococDinr^c;)^'«- cj>cooincDin'*o^inegoooc»r«-co"*ocO"^oococoh-.c3>'«- COOOOOT-h^CRr^OOCO^CMCO-M-^OOCMCMCOCO^O^CROr^^ I^CMCMh^-pTtlpph-r^p^ppp^^CM^ppp-r^^T-'f-; X Iso n nue d dy-y^ddddy^ddddddddddddddy-^y^y^yr^ € 4> S o H O co^eg^in^ooTroocj>cM-f-coo-f-^cocO'^ocMT-inT-cDCD *-*S;Zl2r:5?'''"^*>00T-O0>'«-C0<»C0C3>C3)00T-"«-OC3>^CD r"SSS2S2*^'i5"0^cj>incMcooocoh-'*egcoincooocO"^oo A.I . so n X r>-pppppu)ppp^in^co^'^^in^in^cDO>T-oot^ u 4> J ooy-ddddddddddddddddddddy^dd ti H SH 134 oocMoococooincDcoinoooooo CO CJ> 00 CJ> '^ooa>inincMa>a>a>cooooooo 00 ^ ^ oo C= UH "t^OOCR^^CMOOO-^CMCMOOOOCOCRr^ C3> CM CM ^ ^i^cocopcopppppi^-*^. egpp ^ CO o o i2 X dddddddy^dddddddd 4) , dddd '^OT--T-eocMino^^oooooo r^ -«- '* h* or<>cDoooinoocDh-cj>oooooo T- CD O) h- cMCMinooegcj>ocoo-«-T-cMoocMr>-CM m ^ CO h- p 5 cocMCMCMinini^inoooh>-cDoo^coc3> C3> m o o •S X T-^-r^T-^-t-^T-^-r^-r^t-^CM-r^r-^T^CMCMCMCM y^ d d d H CQ coo)T-r-ooOf-o^cooooooo CD in -r- 00 cDoocoa>cococMr^oocDOooooo 00 r^ 00 m T-cj>oococj>incDcO"*cO'*inco<3>T-co o CO •«- CO o-^-^mcoi^'^h-ocnr^cDcooooo 3 o "^ h-; CM p o P M X T-^ d d d d d d d T-^ d d d •«-: T^ T^ d 4) -^^ dddd P O U COCMCOCMOOm^t-^OOOOOOO CD CM r^ o 00>COCJ>COr-00"*C3)CMOOOOOO CO CJ> CO 1^ CMC0CM00CMinCM"^C0^O^C»CJ>inO m r^ ^ CM O ^ coinin^cDCDcoooa>oooocDin'<(i-^co P T- p o 4> ^ X dddddddddddddddd CJ ^ dddd 135