Proc. Nati. Acad. Sci. USA Vol. 91, pp. 2290-2294, March 1994 Evolution Multiple origins of a spider radiation in Hawaii (Tetragnatha/colonization/morphology/moleuar phylgeny) ROSEMARY G. GILLESPIE*t, HENRIETTA B. CROOMt, AND STEPHEN R. PALUMBI§ *Hawaiian Evolutionary Biology Program and §Department of Zoology, University of Hawaii, HonolulU, HI 96822; and *Department of Biology, University of the South, Sewanee, TN 37375 Communicated by Hampton L. Carson, November 15, 1993 ABSTRACT The Hawaiian Islands are renowned for some patterns and mechanisms of explosive speciation. However, ofthe most spectacular species radiations in the world. Most of a critical element to a comprehensive interpretation of these these radiations have been attributed to single colonization extraordinary radiations is whether the radiation is a conse- events, although the evidence supporting monophyletic origins quence of a single or multiple introductions. Knowledge of is often poorly resolved and/or ambiguous. Without a concrete the number of introductions (i) establishes the phylogenetic understanding of the origins of species radiations, it is impos- status of the species radiation, (ii) shows whether the range sible to understand the phylogenetic pattern of species prolif- ofmorphological and ecological strategies that exist currently eration or the spectrum of morphological, ecological, and arose from a single source or whether part of this variation behavioral modifications attributable to a single colonist. In was "injected" into the community by multiple introduc- this study we examined the species radiation ofthe spider genus tions, and (iii) allows events subsequent to colonization to be Tetragtha in Hawaii. Unlike their mainland congeners, the studied as replicated systems, in situations where closely Hawaiian Tetragnatha are extremely diverse in morphology, related taxa are responsible for multiple colonization events. ecology, and behavior. We tested whether this diversity arose However, the problem of identifying the number of colo- from a single or multiple colonization events. We coupled nists responsible for a given radiation has proved enigmatic. morphological (37 characters) and molecular (sequence from Only recently (50) have the appropriate tools been developed the 12S ribosomal subunit of mitochondrial DNA) approaches to assess the number of separate colonizations within a to assess the phylogenetic position ofthe Hawaiian Tetragnatha lineage by using phylogenetic systematics (10). This ap- relative to continental congeners and to examine evdence for proach has been employed recently to establish the mono- phyly of the Hawaiian Drosophila (Drosophilidae, ref. 11), monophyly. We provide evidence that the Hawaiian Tetra- Ptycta (Psocidae, ref. 12), and silverswords (Asteraceae, ref. gnatha emanate from multiple origins. At least two indepen- 13). Each of these cases provides evidence that a huge dent species radiations, the "spiny-leg" clade and the web- amount of morphological variability can arise during species building species Tetragnatha stelarobusta and Tetragnatha diversification (14). Similar analysis of the Hawaiian Scap- acuta, have arisen from one or more founder events. Two tomyza (Drosophilidae, ref. 11) suggests that the comparable additional natural colonizations have resulted in the establish- diversity in this group is of multiple origin, although the ment of non-speciose lineages, as represented by Tetragnatha evidence is not strong (15). Currently there is no definitive hawaiensis and Doryonychus raptor. evidence for repeated invasions of a speciose genus in Hawaii. Colonization of isolated islands involves both the arrival of The spider genus Tetragnatha is one of the most recent viable propagules and their subsequent establishment. Mac- radiations to have been discovered in the Hawaiian Islands Arthur and Wilson (1) modeled species numbers that would (16, 17). It is one ofonly four spider genera to be substantially be expected on a given island and hypothesized a strong represented in the Hawaiian archipelago (18). The Hawaiian inverse relationship between species richness and distance Tetragnatha are extraordinarily diverse in morphology, ecol- from the source of colonists. The Hawaiian archipelago falls ogy, and behavior, thus presenting a striking contrast to their at the extreme end of this relationship, being the most extremely homogeneous congeners throughout the rest ofthe isolated archipelago in the world. The chain is separated from world (19). Included among the Hawaiian tetragnathids are the nearest continental land mass by nearly 4000 km ofocean groups of species that have relinquished web building and and is noted for its lack of representation by many normally have developed either long leg spines (16) or elongate tarsal prominent biotic components. Before human occupation, the claws (20, 21) for raptorial feeding. Other groups build webs isolation of the islands proved to be a major barrier to in specific habitats uncharacteristic of Tetragnatha, includ- colonization. However, among the small number ofcolonists ing among the foliage and bark of the forest canopy and that reached the archipelago, speciation has occurred repeat- among roots and moss on the forest floor (ref. 17 and edly, often with such extraordinary magnitude that single unpublished data). Taken together, these numerous species colonists have given rise to large numbers of morphologi- dominate the niche of exposed vegetation for terrestrial cally, ecologically, and behaviorally diverse species. The invertebrate predators. Hawaiian Islands are therefore characterized by phenome- In this study we use the radiation of Tetragnatha in Hawaii nally high numbers of species representing very few species to test the popular notion that such radiations tend to be groups (2) and provide some of the best illustrations of such monophyletic. A single introduction would provide further species swarms, including both animals [e.g., drepaniid hon- evidence of the extraordinary diversification possible after eycreepers (3, 4), land snails (5), crickets (6), and Drosophila single founder events (14, 22). Multiple introductions of (7, 8)] and plants [e.g., beggar's ticks and tarweeds (9)]. individuals of the same genus would allow comparison of the Studies ofthese species radiations in the Hawaiian Islands outcome of repeated founder events. We coupled morpho- have demonstrated their enormous potential for elucidating logical and molecular approaches to assess the phylogenetic The publication costs of this article were defrayed in part by page charge Abbreviations: PTP, cladistic permutation tail probability; T-PTP, payment. This article must therefore be hereby marked "advertisement" topology-dependent PTP. in accordance with 18 U.S.C. §1734 solely to indicate this fact. tTo whom reprint requests should be addressed. 2290 Downloaded by guest on October 6, 2021 Evolution: Gillespie et al. Proc. Nadl. Acad. Sci. USA 91 (1994) 2291 position of the Hawaiian Tetragnatha relative to continental RESULTS AND DISCUSSION and to examine evidence for monophyly. congeners Phylogenetic Analysis of Morphological Data. Character changes were weighted equally (compensatory weighting MATERIALS AND METHODS according to number of character states), and a branch and bound search gave three minimum-length (109) trees; con- The genus Tetragnatha is one ofthe largest genera of spiders, sistency index, 0.503, retention index, 0.626 (Fig. 1A). The with -290 described species, excluding synonyms. The ra- trees differed only in their placement of the two Hawaiian diation of Tetragnatha in Hawaii comprises a large number of species T. stelarobusta and T. acuta relative to the T. species, only 25 of which have been described (16-18, 23). waikamoi-T. perreirai group. Subsequent reweighting by We examined representatives of the genus from Hawaii and successive approximations (characters weighted by maxi- from the continents of Asia, Australia, and North America. mum value of the rescaled consistency indices, refs. 30 and The Hawaiian Doryonychus raptor is a highly modified 31) gave two minimum-length trees after a single iteration, species, and the sole representative of the genus is D. raptor with a consensus identical to Fig. 1A; consistency index, from Kauai the 0.717, retention index, 0.814. Bootstrapping (1000 replica- (18, 20, 21). However, gross body morphol- tions) was used (with all character changes equally weighted) ogy, eyes, and genitalic structure of D. raptor are typical of to assess confidence in the analysis. the genus Tetragnatha (H. W. Levi, personal communica- Tests for Nonmonophyly in Morphological Data. In order to tion). Because this species represents extreme specialization evaluate whether the observed hierarchical structure in the and because its relationship to the other Hawaiian spiders is data might represent random character covariation, Faith and obscure, we included it in our analysis. Cranston (32) developed tests of cladistic permutation tail Phylogenetic Analysis of Morphological Data. We scored a probability (PTP) tests. Recent modifications of these tests total of 37 morphological characters relating to the cheliceral have allowed specific phylogenetic topologies to be tested, armature and shape, leg spination and length, ocular pattern, specifically as they relate to monophyly, by using topology- cephalothoracic shape, and shape of the seminal receptacles dependent PTP tests (T-PTP test; ref. 33). We used T-PTP in the female and to the