Wood-feeding cockroaches as models for termite evolution (Insecta: Dictyoptera): Cryptocercus vs. Parasphaeria boleiriana

Available online at www.sciencedirect.com

Molecular Phylogenetics and Evolution 46 (2008) 809–817 www.elsevier.com/locate/ympev

Review

Wood-feeding cockroaches as models for termite evolution (Insecta: Dictyoptera): Cryptocercus vs. Parasphaeria boleiriana Klaus-Dieter Klass a, Christine Nalepa b, Nathan Lo c,* a

State Natural History Collections Dresden, Museum of Zoology, Ko¨nigsbru¨cker Landstrasse 159, D-01109 Dresden, Germany b Department of Entomology, North Carolina State University, Campus Box 7613, Raleigh, NC 27695-7613, USA c Behaviour and Genetics of Social Insects Laboratory, School of Biological Sciences, The University of Sydney, Sydney, NSW 2006, Australia Received 11 July 2007; revised 12 November 2007; accepted 28 November 2007 Available online 15 January 2008

Abstract Isoptera are highly specialized cockroaches and are one of the few eusocial insect lineages. Cryptocercus cockroaches have appeared to many as ideal models for inference on the early evolution of termites, due to their possible phylogenetic relationship and several shared key attributes in life history. Recently, Pellens, Grandcolas, and colleagues have proposed the blaberid cockroach Parasphaeria boleiriana to be an alternative model for the early evolution in termites. We compare the usefulness of Cryptocercus and P. boleiriana as models for termite evolution. Cryptocercus and lower Isoptera (1) can both feed on comparatively recalcitrant wood, (2) have an obligate, rich and unique hypermastigid and oxymonadid fauna in the hindgut, (3) transfer these flagellates to the next generation by anal trophallaxis, (4) have social systems that involve long-lasting biparental care, and, finally, (5) are strongly suggested to be sister groups, so that the key attributes (1)–(4) appear to be homologous between the two taxa. On the other hand, P. boleiriana (1) feeds on soft, ephemeral wood sources, (2) shows no trace of the oxymonadid and hypermastigid hindgut fauna unique to Cryptocercus and lower Isoptera, nor does it have any other demonstrated obligate relationship with hindgut flagellates, (3) is likely to lack anal trophallaxis, (4) has only a short period of uniparental brood care, and (5) is phylogenetically remote from the Cryptocercus + Isoptera clade. These facts would argue against any reasonable usage of P. boleiriana as a model for the early evolution of Isoptera or even of the clade Cryptocercus + Isoptera. Cryptocercus thus remains an appropriate model-taxon-by-homology for early termite evolution. As compared to P. boleiriana, some other Blaberidae (such as the Panesthiinae Salganea) appear more useful as model-taxa-by-homoplasy for the early evolution of the Cryptocercus + Isoptera clade, as their brooding behavior is more elaborate than in P. boleiriana. Ó 2007 Elsevier Inc. All rights reserved. Keywords: Cryptocercus; Isoptera; Protozoa; Eusociality; Biparental care; Xylophagy; Symbionts

1. Introduction Termites (Blattodea1: Isoptera) are highly specialized cockroaches within the superorder Dictyoptera (Kristen*

Corresponding author. E-mail addresses: [email protected] (K.-D. Klass), [email protected] (C. Nalepa), [email protected] (N. Lo). 1 The terms ‘‘Blattodea” and ‘‘Blattaria” have been variously used for cockroaches under exclusion (more usually) or inclusion of termites. We follow Hennig (1981): Blattaria is the assemblage comprising all cockroaches but not termites, while Blattodea comprises both cockroaches and termites. With Isoptera being the sister group of the cockroach genus Cryptocercus, only Blattodea is a monophyletic taxon, but not Blattaria. 1055-7903/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2007.11.028

sen, 1991; Klass, 2003; Terry and Whiting, 2005; Kjer et al., 2006; Inward et al., 2007; Lo et al., 2007b). They represent one of the few insect lineages that show a highly elaborate social organization. The origin of isopteran eusociality has long been of considerable interest to entomologists, and its discussion is necessarily linked to discussions on the phylogeny of Dictyoptera. In recent years there has been overwhelming morphological (Klass, 1995; Deitz et al., 2003; Klass and Meier, 2006) and molecular evidence (Lo et al., 2000, 2003, 2007a; Terry and Whiting, 2005; Kjer et al., 2006; Pellens et al., 2007; Inward et al., 2007) that Isoptera is the sister group of the cockroach genus Cryptocercus and thus a subordinate clade of Blattodea.

810

K.-D. Klass et al. / Molecular Phylogenetics and Evolution 46 (2008) 809–817

This is in strong contrast to earlier hypotheses that assumed a position of Isoptera outside a monophyletic Blattaria1 (e.g., Thorne and Carpenter, 1992; Kambhampati, 1995) and a subordinate position of Cryptocercus inside the cockroach family Polyphagidae (Grandcolas, 1994, 1997a). Cryptocercus cockroaches have appeared to many as ideal models for inference on the early evolution of termites, due not only to their possible close phylogenetic relationship, but also to several key attributes shared between Cryptocercus and ‘‘lower Isoptera” (=Isoptera excluding the most derived and species-rich family Termitidae) (e.g., Honigberg, 1969; Bobyleva, 1975; McKittrick, 1965; Cleveland et al., 1934). (1) Both are xylophagous. (2) Both have an obligate relationship with numerous species of the flagellate protistan taxa Hypermastigida and Oxymonadida in their hindguts, with many flagellate genera co-occurring in Cryptocercus and subgroups of Isoptera. (3) Both transfer these flagellates to the next generation by anal trophallaxis (=proctodeal feeding). (4) Both show advanced social behaviors: Cryptocercus lives in biparental family groups (subsociality) (Seelinger and Seelinger, 1983; Nalepa, 1984; Nalepa et al., 1997, 2001a; Park et al., 2002). Biparental subsociality is also seen in dealate pairs of Isoptera during the early stages of colony foundation (e.g., Noirot, 1985; Roisin, 1990); the family switches to alloparental care (eusociality) with the appearance of workers or pseudergates. These four features are intimately correlated because xylophagy in Cryptocercus and lower termites partly depends on the contribution of the flagellates to digestion (for host-flagellate relationships see Cleveland, 1924, 1925; Yamin and Trager, 1979; Bignell, 2000; Inoue et al., 2000), and the mechanism of transferring these flagellates between generations, anal trophallaxis, is related to both the degree of host-microbe interdependence and to host social structure (Nalepa et al., 2001b). In recent years Pellens, Grandcolas, and colleagues have been studying in detail the cockroach Parasphaeria boleiriana from Brazil (Pellens et al., 2002; Brugerolle et al., 2003). This species belongs to the Blaberidae, a family that is, without question, distantly related to Cryptocercus and Isoptera. Recently, Pellens et al. (2007) suggested that P. boleiriana is a valuable alternative model for the early evolution of termites, based on the surmised presence of three of the four above-mentioned key attributes in P. boleiriana: hindgut flagellates, brood care, and xylophagy; the mode of flagellate transmission to the progeny remains unknown. Pellens et al. (2007) furthermore consider the phylogenetic relationship between Cryptocercus and Isoptera doubtful, which would decrease the value of Cryptocercus as a model taxon for termite evolution. We disagree with many of the reflections and conclusions in Pellens et al. (2007) and here put our concerns forward for further discussion. Three issues are important: (1) whether Cryptocercus can be regarded as a model for early evolution of termites based on phylogenetic relationship (and homology of key attributes); (2) to what extent there

are actually striking similarities between P. boleiriana and (lower) Isoptera in the key attributes cited by Pellens et al. (2007), i.e., (2a) hindgut flagellates, (2b) brood care and sociality, and (2c) xylophagy; (3) to what extent a phylogenetically distant taxon can be useful as a supplementary model for the early evolution of termites (based on parallel evolution of key attributes). 2. Discussion (1) Phylogenetic relationships: Cryptocercus + Isoptera or Cryptocercus + Polyphagidae? Our first concern relates to Pellens et al.’s (2007) assessment of the state-of-the-art of phylogeny reconstruction in Dictyoptera, regarding Cryptocercus, Isoptera, and the cockroach family Polyphagidae in particular. Pellens et al. (2007) make a very tentative statement that Cryptocercus and Isoptera might together form a clade, citing Lo et al. (2000) (and Maekawa et al., 2005a, which, however, has no bearing on this issue, as it deals only with the internal phylogeny of Cryptocercus). Indeed, all phylogenetic evidence of recent years strongly supports this hypothesis. Support comes from the morphological side by Klass and Meier’s (2006) cladistic analysis of 175 characters for 27 dictyopteran taxa (see also Deitz et al., 2003; Klass, 1995). In that study a clade Cryptocercus + Isoptera was obtained both including and excluding the characters referring to oxymonadid and hypermastigid flagellates (characters 157, 158, and analyses RC in Table 4 therein). The molecular analyses by Lo et al. (2003: 12S, 16S, 18S, and COII, and 16S of Blattabacterium endosymbionts) also clearly recovered the Cryptocercus + Isoptera clade, as did the studies of Kjer et al. (2006: 8 genes), Terry and Whiting (2005: 18S, 28S, and H3), and Lo et al. (2007a: 18S, COII, and H3). Importantly, this grouping was recovered using a variety of different analytical treatments. A molecular analysis of 12S, 18S, 28S, COII, and H3 for the most inclusive taxon sample so far (107 species of Dictyoptera plus 11 from the outgroup) has further confirmed the Cryptocercus + Isoptera clade (Inward et al., 2007), and this clade is also strongly supported in Pellens et al.’s (2007, Fig. 1) own analysis of the genes 12S, 16S, 18S, and COII (Jackknife value 94%, Bremer support 16). The Cryptocercus + Isoptera clade thus receives high support from each of several independent data sources. Pellens et al. (2007) note that Kambhampati (1995: 12S and 16S) and Maekawa and Matsumoto (2000: COII) show Cryptocercus as a ‘‘particular family” within the cockroaches, as evidence suggesting that the Cryptocercus–Isoptera relationship is not well secured. However, this is not valid for Maekawa and Matsumoto (2000), who a priori defined Isoptera as the outgroup to Blattaria; the position of the root in their trees, and hence the relationship between Cryptocercus and Isoptera based on their data set, can only be found by a re-analysis using non-dictyopteran outgroup taxa. Concerning Kambhampati (1995), we note that Grandcolas and D’Haese (2001) have demonstrated serious

K.-D. Klass et al. / Molecular Phylogenetics and Evolution 46 (2008) 809–817

shortcomings for the treatment of DNA data in Kambhampati (1996: only 12S), and much of this also applies to Kambhampati (1995: 12S and 16S). Furthermore, all genes in question—COII, 12S, and 16S—are also included in the combined analysis of four genes in Pellens et al. (2007), which eventually strongly supports the Cryptocercus + Isoptera clade. The only evidence that may still contradict a Cryptocercus + Isoptera clade is the analysis of hypertrehalosaemic neuropeptides in Ga¨de et al. (1997), but this is based on very few informative characters and shows little phylogenetic resolution. The morphological evidence on the phylogenetic position of Cryptocercus has long been disputed between K.-D. Klass (1995, 1997, 1998, 2001; Klass and Meier, 2006), who considers this genus sister to Isoptera, and Grandcolas (1994, 1996, 1997a,b, 1999), who finds Cryptocercus deeply nested in the Polyphagidae. The numerous shortcomings in the morphological work of Grandcolas (1994, 1996) were discussed in detail in Klass (1997: 327–337; 2001). In addition, Grandcolas (1996) used Isoptera as an outgroup taxon for his cladistic analysis of Blattaria and thus could not test the relationship between Isoptera and Cryptocercus. Pellens et al. (2007) also cite Grandcolas et al. (2001) as evidence of the Cryptocercus–polyphagid relationship; however, this molecular study includes only members of Cryptocercus, the polyphagid Therea petiveriana, and the blattid Blatta orientalis, which is defined as the outgroup. Pellens et al. (2007) claim that ‘‘we do not intend to solve the old-standing controversy of the Cryptocercus position which would need a specific taxon sampling with ‘basal’ cockroach groups such as the family Polyphagidae”. This is odd, because a polyphagid is actually included in the tree in Pellens et al. (2007, Figs. 1–3): Therea petiveriana, which according to Grandcolas (1994) is the polyphagid most closely related to Cryptocercus. Another polyphagid, Polyphaga aegyptiaca, is included in the multiple-gene analysis of Kjer et al. (2006), and both polyphagid species have been sampled for Lo et al.’s (2003) analyses. In neither of these studies do the polyphagids cluster with Cryptocercus. Regarding the most recent studies, Lo et al. (2007a) included Ergaula capucina, T. petiveriana, and P. aegyptiaca, and Inward et al. (2007) included these plus four further Polyphagidae from different subfamilies. In both analyses all Polyphagidae clustered together unambiguously, while the polyphagid clade is remote from the Cryptocercus + Isoptera clade. Consequently, the hypothesis of Cryptocercus being nested in Polyphagidae should eventually be dismissed. The strong support for a Cryptocercus + Isoptera clade means that all similarities between Cryptocercus and Isoptera (or the paraphyletic ‘‘lower Isoptera”) are most parsimoniously interpreted as homologies, including aspects of xylophagy, biparental care, anal trophallaxis, and the rich oxymonadid–hypermastigid hindgut ‘fauna’. Differences between Cryptocercus and Isoptera appear as autapomorphies of either taxon; some of the respective characters can be polarized by outgroup comparison with other cockroaches.

811

(2a) Hindgut flagellate fauna. Cryptocercus and lower Isoptera share many oxymonadid and hypermastigid taxa in their hindgut that are unique to them (indeed, among all animals!), such as members of Spirotrichosomidae, Hoplonymphidae, Staurojoeninidae, and Eucomonymphidae. Finding any of these flagellate taxa in a distantly related cockroach would really be striking. Pellens et al. (2007) state that P. boleiriana ‘‘permanently harbours intestinal flagellates” and consider this a key attribute shared with lower termites (and Cryptocercus). Clearly, most subgroups of Oxymonadida and Hypermastigida have so far only been found in hindguts of Cryptocercus and lower Isoptera. However, it is well known that a few species of both flagellate taxa show a much wider occurrence; they are found in cockroaches other than Cryptocercus as well as some other insects (e.g., Lorenc, 1939; Roth, 1982; Parker, 1982; Mo¨hn, 1984, 171). In addition, there are flagellates from other major taxa (e.g., Trichomonadida, Diplomonadida) that occur in the guts of many insects and other animals. As already noted by Klass and Meier (2006, 23), it is thus highly important to have a detailed look at the intestinal flagellates of P. boleiriana. Pellens et al. (2002) reported a variety of trichomonadid species for P. boleiriana. Brugerolle et al. (2003) further specified these as members of the genera Monocercomonas and Tetratrichomastix; they additionally reported the presence in P. boleiriana of two oxymonadid species of the genera Monocercomonoides and Polymastix and one diplomonadid species resembling Hexamita. The trichomonadids Monocercomonas and Tetratrichomastix also occur in a variety of insects including cockroaches other than Cryptocercus. Monocercomonoides and Polymastix are two basal oxymonadid (s.l.) genera that are widely distributed in the guts of cockroaches and other terrestrial arthropods. Hexamita includes many free-living species, as well as species that are parasitic or commensal in a variety of metazoan groups (for life history data of flagellates see, e.g., Mo¨hn, 1984, 166, 171; Parker, 1982, 500; Brugerolle et al., 2003). Consequently, for P. boleiriana not a single one of the many flagellate taxa that are unique to Cryptocercus and lower Isoptera has been demonstrated. The results on P. boleiriana thus instead underline the uniqueness of the gut fauna shared between Cryptocercus and Isoptera, and they do not at all support the proposal of P. boleiriana being useful as a model for termite evolution. Pellens et al. (2007) claim that the presence of hindgut flagellates distinguishes P. boleiriana from another taxon of subsocial, xylophagous cockroaches, the Panesthiinae (Blaberidae; including, among others, the genera Panesthia and Salganea; see Fig. 1 in Pellens et al., 2007). Kidder (1937), however, reported several flagellates from Panesthia javanica, including the genera Monocercomonoides and Hexamita, which are among those reported for P. boleiriana. The shared possession of these flagellates in Panesthiinae and in P. boleiriana was noted by Brugerolle et al. (2003).

812

K.-D. Klass et al. / Molecular Phylogenetics and Evolution 46 (2008) 809–817

We note that, regardless of their diet, all examined cockroaches have hindguts that are fermentation tanks filled with a diverse anaerobic microbiota, including ciliates, amoebae, flagellates, and a heterogeneous prokaryotic assemblage (Kidder, 1937; Steinhaus, 1946; Guthrie and Tindall, 1968; Bracke et al., 1979; Bignell, 1981; Cruden and Markovetz, 1984; Sanchez et al., 1994; Zurek and Keddie, 1996; Lilburn et al., 2001). Parasphaeria boleiriana is not unique in this respect. Furthermore, unlike the case for Cryptocercus/lower termites, there is currently no evidence that P. boleiriana has an obligate relationship with any of its described gut flagellates. Brugerolle et al. (2003) indicate that these flagellates occur variably, they are probably not host specific, and that their food vacuoles do not contain wood particles. This further decreases the value of P. boleiriana as a model for the early evolution of termites. (2b) Brood care, sociality, and anal trophallaxis. Brood care in Cryptocercus is strictly biparental and lasts until the death of adults; this is typically three or so years in the field, with females outliving males (Nalepa, 1984; Nalepa et al., 1997, 2001a; Park et al., 2002; Seelinger and Seelinger, 1983). Biparental care is also found in the Isoptera during the early stages of colony foundation (e.g., Noirot, 1985; Roisin, 1990). In both Cryptocercus and lower Isoptera, adults feed neonates on hindgut fluids (anal trophallaxis). This behavior also assures the intergenerational transmission of flagellates. Pellens et al. (2007) refer to the social structure of P. boleiriana as ‘‘brood care”, and Pellens et al. (2002, 353) indicate that it is ‘‘undoubtedly convergent” with that of Cryptocercus. But does brood care in P. boleiriana include striking similarities with termites and Cryptocercus, and could it include behavior aimed at transmitting the gut fauna? Brood care and social structure in P. boleiriana is much less elaborate than in Cryptocercus and incipient termite colonies, and similar to that known from some other cockroaches. According to Pellens et al. (2002, 352), by far the largest social class of P. boleiriana in the field (54%) consisted of nymphs found together, without adults. The description given in the text is ‘‘Most colonies consisted of small to large groups of nymphs, sometimes with a few adults”. Only 4.3% of the social groups consisted of adult males and females together with nymphs, and it is unclear as to how many adults of each sex were present in these groups. Females of P. boleiriana that gave birth in the laboratory did exhibit brooding behavior for about 12 days before the offspring dispersed (Pellens et al., 2002, 353); nymphs remained in close proximity or in physical contact with their mother during this time, but no other interactions were observed. ‘‘Parasphaeria boleiriana can thus be characterized as a gregarious species with a relatively short period of brood care” (Pellens et al., 2002, 356). Due to the unspecific and likely non-obligate relationship between P. boleiriana and its gut flagellates (see above), some refined mechanism for intergenerational transmission of the gut fauna (such as anal trophallaxis)

is not expected. Until demonstrated otherwise, the most likely scenario is that neonates acquire their complement of gut fauna by eating the feces of conspecifics, by ingesting the microbes along with their food source, or both—as is the case in many cockroaches (Nalepa et al., 2001b). The social structure of P. boleiriana resembles that of a number of other Blaberidae (Nalepa and Bell, 1997). Byrsotria fumigata, for example, is a gregarious species in which first instar nymphs recognize their own mother and prefer to aggregate beneath her for the first 15 days after hatch (Liechti and Bell, 1975). Several blaberid and one blattellid cockroach are known to exhibit more elaborate (uni- or bi-) parental care than P. boleiriana and B. fumigata, including at least eight species that feed neonates on bodily secretions originating from the stomodeum, the brood sac, the hemolymph, or from tergal or sternal glands (Bell et al., 2007). One example is the wood-feeding blaberid genus Salganea (Panesthiinae), whose members live in biparental families (Maekawa et al., 1999, 2005b; Matsumoto, 1987); in S. taiwanensis, nymphs cling to the mouthparts of their parents and take liquids via stomodeal feeding (Bell et al., 2007: Fig. 8.3; details of the gut microbiota are apparently unknown). Parental care in Dictyoptera that includes proctodeal feeding, however, is known only in Cryptocercus and Isoptera, and apart from Salganea, the same is true for long term biparental care. Pellens et al. (2002, 356) conclude that social structure in P. boleiriana ‘‘cannot be equated with that of a typically subsocial species whose brood care extends over considerable periods or includes special behaviours such as trophallaxis”. We agree. Thus, also with regard to brood care, Parasphaeria boleiriana is much less useful as a model for termite evolution than Cryptocercus or even some other blaberid cockroaches. Furthermore, we find it necessary to comment upon some other statements and assumptions made by Pellens et al. (2007). First, they state without reference that previously it had been assumed that the need for flagellate transmission to the next generation has led to a long and complex brood care in the ancestors of termites (or termites + Cryptocercus), and that the example of P. boleiriana shows that this is not necessarily required. However, since P. boleiriana lacks those flagellates unique to Cryptocercus and Isoptera and may not have an obligate relationship with the flagellates in its hindgut, such inference from conditions in P. boleiriana on those in the early evolution of termites is inappropriate. Second, we disagree with Pellens et al.’s (2007) view of the ‘‘shift-in-dependent-care hypothesis”. The crucial difference in the sociality of Cryptocercus and termites is that only in the latter, after a phase of biparental care, is the care and feeding of young brood taken over by older brood in the family, and hence shifts from parents to older siblings. The ‘‘shift-in-dependent-care hypothesis” (=‘‘trophic shift hypothesis”: Bell et al., 2007) proposes that alloparental care in termites is derived from biparental care in a Cryptocercus-like ancestor (Nalepa, 1988a, 1994). Pellens

K.-D. Klass et al. / Molecular Phylogenetics and Evolution 46 (2008) 809–817

et al. (2007) state that the ‘‘main problem” with using Cryptocercus as a model for the termite ancestor is that ‘‘it shows behavioural traits incompatible with the ‘shiftin-dependent-care hypothesis’: the first brood remains with the female during several years but prevents it to have a second brood, preventing any shift in care from female to broods”. This interpretation is based on a misconception of this verbal model, in which the prevention of reproduction by the offspring is rather considered the driving force (Nalepa, 1988a, 1994; Bell et al., 2007, Fig. 9.8). If the cost of parental care constrained female reproduction in a termite ancestor, then the assumption of that cost by the oldest nymphs of the initial cohort (alloparental care) could free her resources for additional bouts of oviposition. Furthermore, Cryptocercus punctulatus lays up to four oothecae during its sole reproductive bout. There is a lag of 2– 6 days between successive oothecae in the laboratory, and, unlike other oviparous cockroaches, egg hatch within an individual ootheca is staggered in time. Together, these circumstances effectuate age differentials within the brood, with a lag of two or more weeks between the first- and lasthatched nymphs in large clutches (Nalepa, 1988b, 1990, and unpubl. data; Bell et al., 2007, 161). As a result of asynchronous hatching and the quick growth of neonates, both trophically dependent (first and second instars) and trophically independent nymphs (third instars and subsequent) can be contemporaneous in young families. Cryptocercus, then, exhibits both key requirements proposed for a termite ancestor in the trophic shift model: first, parental care that constrains future reproduction, and second, the potential— unrealized in Cryptocercus—to shift that physiological cost to the oldest nymphs in the cohort if the first-hatched nymphs begin taking care of their younger siblings. (2c) Xylophagy. Cryptocercus and lower termites typically live in and ingest fairly recalcitrant dead wood; they use softwood (i.e., coniferous) as well as hardwood logs that may take decades to degrade. Parasphaeria boleiriana clearly ingests the wood of the boleira tree; small wood pieces could be identified in both gut contents and feces (Pellens et al., 2002). Xylophagy in a wide sense is thus actually shared between P. boleiriana and Cryptocercus/ lower termites. However, under closer inspection, this is neither a very specific dietary trait, nor an unusual one among cockroaches. Two things are important to note: first, all studied cockroach and termite species have endogenous cellulase genes, which are known to be transcribed (Lo et al., 2000; Tokuda et al., 2004). This suggests there is a widespread ability among these insects to use cellulose-based materials as food, if only as a supplement to other resources. Only Cryptocercus and lower termites, however, have an additional specific type of cellulose digestion that involves oxymonadid and hypermastigid flagellates requiring vertical intergenerational transmission. Second, rotting wood is a tremendously diverse resource that varies with plant taxon, size, location, degree and type of rot, orientation, presence of other invertebrates, and other factors. The category

813

‘‘xylophagous” is more fluid than generally recognized, and throughout the dietary continuum of rotted leaf litter, soft rotted wood, and hard, freshly-dead wood, divisions are not always easy to make (Bell et al., 2007). Cockroach species from most families have been collected from rotting logs, but in the majority of cases it is unknown whether these feed on wood and associated microbes, if they depart to forage elsewhere, or both. This is particularly true of the many cockroaches that bore into soft wood, such as the rotted trunks of coconut and banana palms. One example is Panchlora nivea, represented in Pellens et al.’s (2007) phylogenetic tree (therein not indicated as ‘‘wood-eating”): nymphs in this genus can be reared to adulthood fed only on the rotting trunks of palms (Wolcott, 1948). Boleira also fits this category of soft, quickly rotting plant litter; Pellens et al. (2007) indicate that half of the trunks they monitored decomposed within two years. While wood is clearly a food source for P. boleiriana, it remains open as to what extent other items may contribute to its diet, at least during certain developmental stages. During their night surveys, Pellens et al. (2002) noted ‘‘thirty individuals, mostly old nymphs. . .moving over dead trunks or on the soil”. Parasphaeria boleiriana, like a variety of cockroach species, may be best described as saproxylic.2 Additional cockroaches in this category include Anamesia douglasi (Blattidae: Polyzosteriinae) (Roach and Rentz, 1998), Methana parva (Blattidae sensu Klass and Meier, 20063), Litopeltis bispinosa (Blaberidae: Epilamprinae) (Roth and Willis, 1960), Salganea spp. and other panesthiine taxa (Blaberidae: Panesthiinae) (e.g., Roth, 1979; Rugg, 1987), Paramuzoa alsopi (Blattellidae: Nyctiborinae) (Grandcolas, 1993), Lauraesilpha mearetoi (Tryonicidae sensu Klass and Meier, 2006) (Grandcolas, 1997b), and Polyphagoides cantrelli (Polyphagidae: Polyphaginae) (Roach and Rentz, 1998). We find it curious that P. alsopi and L. mearetoi, previously reported as xylophagous by one of the authors of Pellens et al. (2007), did not merit discussion in their paper. The fact that P. boleriana, like a number of other cockroaches, lives in an ephemeral wood source (decaying completely within 2–3 years), while Cryptocercus and lower termites nest in wood sources that may take decades to degrade, further weakens the utility of P. boleriana as a model-taxon-by-homoplasy for termites. In Cryptocercus and lower termites the level of recalcitrance of the wood in which they nest is likely to have major implications for colony stability, which in turn may influence the length of brood care and the extent of sibling interactions. In this 2

An organism that depends, during some part of its life cycle, upon dead or dying wood or moribund or dead trees (standing or fallen), or upon wood-inhabiting fungi, or upon the presence of other saproxylics (Speight, 1989). 3 The genus Methana has been assigned to Blattidae–Tryonicinae by Grandcolas (1997b), but the arguments for this placement were rejected in Klass (2001). This genus is thus not included in the Tryonicidae (elevated to family rank) sensu Klass and Meier (2006), which only comprises the genera Tryonicus and Lauraesilpha.

814

K.-D. Klass et al. / Molecular Phylogenetics and Evolution 46 (2008) 809–817

respect, any model-taxon-by-homoplasy for termites should preferably nest in relatively recalcitrant wood. In panesthiine blaberids host choice is similar to that of Cryptocercus: Panesthia cribrata in Australia (Rugg, 1987), as well as species of Panesthia and Salganea in Japan (Maekawa et al., 2005b; K. Maekawa, pers. comm. to CAN) utilize softwood as well as hardwood logs. They generally use what is available, and when population density is high, they are found in a greater variety of log types (D. Rugg, pers. comm. to CAN). Altogether, a shared diet of rotting, cellulose-based material in P. boleiriana and Cryptocercus/lower termites does not constitute a striking similarity and thus further degrades P. boleiriana as a useful model for termite evolution. (3) Usefulness of a phylogenetically distant model for termite evolution. Wood-feeding, biparental care, anal trophallaxis, and the rich oxymonadid/hypermastigid hindgut ‘fauna’ are most parsimoniously considered homologous between Cryptocercus and lower Isoptera due to their highly probable sister group relationship. Thus, Cryptocercus is an ideal model-taxon-by-homology for early termite evolution. The life history of Cryptocercus can be assumed to closely reflect an early evolutionary stage of the termite clade—unless contradictory evidence can be demonstrated for individual aspects of Cryptocercus biology that might be autapomorphic. One probable autapomorphy in Cryptocercus is winglessness and non-aerial dispersal; it seems unlikely that termites re-evolved wings from a wingless ancestor rather than vice versa. Nonetheless, wing reduction may be possible without elimination of the genetic background for macropterous development (Masaki and Shimizu, 1995), and potential evolutionary reversal of wing loss has been proposed to occur in Heteroptera (Andersen, 1997) and Phasmatodea (Whiting et al., 2003; but see Trueman et al., 2004). In what way could an additional, phylogenetically distant model-taxon-by-homoplasy be useful? We advocate that in general any model-taxon-by-homoplasy should only be used in the attempt to understand evolutionary steps that remain unexplained by an available model-taxon-byhomology. Regarding the case of termite evolution, this concerns: (A) how biparental care, obligate association with Oxymonadida/Hypermastigida, and anal trophallaxis evolved in the stem lineage of Cryptocercus + Isoptera; (B) how care for the young brood shifted from parents to siblings (initial stage of eusociality) and (C) how the differentiation into castes (elaboration of eusociality) proceeded in the stem lineage of Isoptera after its divergence from Cryptocercus. As discussed above, the evolutionary state of biparental care, association with Oxymonadida/Hypermastigida, and anal trophallaxis in the early stem lineage of Isoptera is best concluded from conditions in Cryptocercus. Questions related to (B) and (C) need model taxa that show some transition to eusociality. These are not available from cockroaches. Such taxa do exist in Hymenoptera; however, caution is advisable, because the preconditions

in these Hymenoptera are fundamentally different from those in Isoptera (male haploidy, holometaboly, uniparental care, origination from parasitic ancestors, lack of obligate association with Oxymonadida and Hypermastigida). Questions related to (A) need model taxa that show some tendency towards parental care (preferably biparental), association with specific protists (preferably obligate) including their targeted transmission to the progeny (preferably by anal trophallaxis), and a cellulose based diet (preferably using recalcitrant wood). We are unaware of any subgroup of Blattodea other than Cryptocercus and termites that have been shown to depend on and vertically transmit specific, obligate protists. It is thus the criteria of parental care and xylophagy that remain to be assessed. As explained above, P. boleiriana, like several other cockroaches, shows a short period of uniparental care and uses an ephemeral wood source. Regarding these two criteria, P. boleiriana only to a very limited extent approaches the Cryptocercus + Isoptera clade, while there are several more logical choices. One is the blaberid genus Salganea, whose members feed within logs similar to those utilized by Cryptocercus (K. Maekawa, pers. comm.), and live in biparental families. Brood care in the one examined species, S. taiwanensis, includes parental feeding (see above). We emphasize that taxa of this kind cannot act as models for the early evolution of Isoptera; they can serve only as model-taxaby-homoplasy for the early evolution of the Cryptocercus + Isoptera clade! Thus, compared with Cryptocercus, other cockroaches are, at most, of limited utility for conclusions on early termite evolution. For the early evolution of the Cryptocercus + Isoptera clade, there are more useful model-taxaby-homoplasy among cockroaches than P. boleiriana. 3. Conclusions Cryptocercus and lower Isoptera can both feed on comparatively recalcitrant wood, have an obligate, rich, and unique hypermastigid and oxymonadid fauna in the hindgut, transfer these flagellates to the next generation by anal trophallaxis, and their sociality includes long-lasting biparental care (in termites during colony foundation). Based on the strong support for a sister group relationship between Cryptocercus and Isoptera, these key attributes are most parsimoniously considered homologous between the two taxa. The subsocial Cryptocercus thus appears as an ideal model-taxon-by-homology for the early evolution of eusocial Isoptera. Pellens et al. (2007) suggest the blaberid P. boleiriana, which is evidently phylogenetically remote from the Cryptocercus + Isoptera clade, to be a useful alternative model-taxon for the early evolution of termites. However, it shows no trace of the unique oxymonadid and hypermastigid hindgut fauna common to Cryptocercus and lower Isoptera; neither is there evidence for obligate relationships with any other hindgut flagellates. It shows only a short period of uniparental (instead of strictly biparental) brood

K.-D. Klass et al. / Molecular Phylogenetics and Evolution 46 (2008) 809–817

care. Anal trophallaxis (or other feeding interactions) during brood care has not been observed, and is unlikely to be present. Short term brooding behavior by females within the larger context of gregarious behavior is not an unusual social structure in Blaberidae. Xylophagy is shared with Cryptocercus and lower termites, but in P. boleiriana it relates to soft, ephemeral wood, while wood-feeding of this type may not be as rare in cockroaches as Pellens et al. (2007) assume. These facts would argue against any reasonable usage of P. boleiriana as a model for the early evolution of Isoptera or even of the clade Cryptocercus + Isoptera. On the other hand, some other members of the Blaberidae (Panesthiinae such as Salganea species) appear more useful as model-taxa-by-homoplasy for the early evolution of the Cryptocercus + Isoptera clade. With these conclusions we are in sharp contrast to Pellens et al. (2007). References Andersen, N.M., 1997. Phylogenetic tests of evolutionary scenarios: the evolution of flightlessness and wing polymorphism in insects. Me´m. Mus. Nat. Hist. Nat. 173, 91–108. Bell, W.J., Roth, L.M., Nalepa, C.A., 2007. Cockroaches: Ecology Behavior and Natural History. The Johns Hopkins University Press, Baltimore, pp. 230. Bignell, D.E., 1981. Nutrition and digestion. In: Bell, W.J., Adiyodi, K.G. (Eds.), The American Cockroach. Chapman and Hall, New York, pp. 57–86. Bignell, D.E., 2000. Introduction to symbiosis. In: Abe, T., Bignell, D.E., Higashi, M. (Eds.), Termites: Evolution, Sociality, Symbioses, Ecology. Kluwer Academic Publishers, Dordrecht, Netherlands, pp. 189– 208. Bobyleva, N.N., 1975. Morphology and evolution of intestinal parasitic flagellates of the far-eastern roach Cryptocercus relictus. Acta Protozoologica 14, 109–160. Bracke, J.W., Cruden, D.L., Markovetz, A.J., 1979. Intestinal microbial flora of the American cockroach, Periplaneta americana L.. Appl. Environ. Microbiol. 38, 945–955. Brugerolle, G., Silva-Neto, I.D., Pellens, R., Grandcolas, P., 2003. Electron microscopic identification of the intestinal protozoan flagellates of the xylophagous cockroach Parasphaeria boleiriana from Brazil. Parasitol. Res. 90, 249–256. Cleveland, L.R., 1924. The physiological and symbiotic relationships between the intestinal Protozoa of termites and their host, with special reference to Reticulitermes flavipes Kollar. Biol. Bull. 46, 178–227. Cleveland, L.R., 1925. The effects of oxygenation and starvation on the symbiosis between the termite, Termopsis, and its intestinal flagellates. Biol. Bull. 48, 309–327. Cleveland, L.R., Hall, S.R., Saunders, E.P., Collier, J., 1934. The woodfeeding roach Cryptocercus, its Protozoa, and the symbiosis between Protozoa and roach. Mem. Am. Acad. Sci. 17, 185–342. Cruden, D.L., Markovetz, A.J., 1984. Microbial aspects of the cockroach hindgut. Arch. Microbiol. 138, 131–139. Deitz, L.L., Nalepa, C.A., Klass, K.-D., 2003. Phylogeny of the Dictyoptera re-examined. Entomol. Abh. 61, 69–91. Ga¨de, G., Grandcolas, P., Kellner, R., 1997. Structural data on hypertrehalosaemic neuropeptides from Cryptocercus punctulatus and Therea petiveriana: how do they fit into the phylogeny of cockroaches? Proc. R. Soc. Lond. B 264, 763–768. Grandcolas, P., 1993. Le genre Paramuzoa Roth, 1973: sa re´partition et un cas de xylophagie chez les Nyctiboriniae (Dictyoptera, Blattaria). Bull. Soc. Entomol. Fr. 98, 131–138.

815

Grandcolas, P., 1994. Phylogenetic systematics of the subfamily Polyphaginae, with the assignment of Cryptocercus Scudder, 1862 to this taxon (Blattaria, Blaberoidea, Polyphagidae). Syst. Entomol. 19, 145–158. Grandcolas, P., 1996. The phylogeny of cockroach families: a cladistic appraisal of morpho-anatomical data. Can. J. Zool. 74, 508–527. Grandcolas, P., 1997a. What did the ancestors of the woodroach Cryptocercus look like? A phylogenetic study of the origin of subsociality in the subfamily Polyphaginae (Dictyoptera, Blattaria). In: Grandcolas, P. (Ed.), The Origin of Biodiversity in Insects: Phylogenetic Tests of Evolutionary scenarios. Me´m. Mus. Nat. Hist. Nat. 173, 231–252. Grandcolas, P., 1997b. Syste´matique phyloge´ne´tique de la sous-famille des Tryonicinae (Dictyoptera, Blattaria, Blattidae). In: Nait, J., Matile, L. (Eds.), Zoologia Neocaledonica, vol. 4. Me´m. Mus. Nat. Hist. Natur. 171, 91–124. Grandcolas, P., 1999. Reconstructing the past of Cryptocercus (Blattaria: Polyphagidae): phylogenetic histories and stories. Ann. Entomol. Soc. Am. 92, 303–307. Grandcolas, P., D’Haese, C., 2001. The phylogeny of cockroach families: is the current molecular hypothesis robust? Cladistics 17, 48–55. Grandcolas, P., Park, Y.C., Choe, J.C., Piulachs, M.-D., Belle´s, X., D’Haese, C., Farine, J.-P., Brossut, R., 2001. What does Cryptocercus kyebangensis, n.sp. (Dictyoptera, Blattaria, Polyphagidae) from Korea reveal about Cryptocercus evolution? A study in morphology, molecular phylogeny, and chemistry of tergal glands. Proc. Acad. Nat. Sci. Phila. 151, 61–79. Guthrie, D.M., Tindall, A.R., 1968. The Biology of the Cockroach. Edward Arnold Ltd., London, 408 pp. Hennig, W., 1981. Insect Phylogeny. Wiley Interscience Publication, John Wiley & Sons, Chichester, etc. Honigberg, B.M., 1969. Protozoa associated with termites and their role in digestion. In: Krishna, K., Weesner, F.M. (Eds.), The Biology of Termites, vol. 2. Academic Press, New York, pp. 1–36. Inoue, T., Kitade, O., Yoshimura, T., Yamaoka, I., 2000. Symbiotic associations with protists. In: Abe, T., Bignell, D.E., Higashi, M. (Eds.), Termites: Evolution, Eusociality, Symbioses, Ecology. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 275–288. Inward, D., Beccaloni, G., Eggleton, P., 2007. Death of an order: a comprehensive molecular phylogenetic study confirms that termites are eusocial cockroaches. Biol. Lett. 3, 327–330. doi:10.1098/ rsbl.2006.061. Kambhampati, S., 1995. A phylogeny of cockroaches and related insects based on DNA sequence of mitochondrial ribosomal RNA genes. Proc. Natl. Acad. Sci. USA 92, 2017–2020. Kambhampati, S., 1996. Phylogenetic relationship among cockroach families inferred from mitochondrial 12S rRNA gene sequence. Syst. Entomol. 21, 89–98. Kidder, G.W., 1937. The intestinal Protozoa of the wood-feeding roach Panesthia. Parasitology 29, 163–203, 10 plates. Kjer, K.M., Carle, F.L., Litman, J., Ware, J., 2006. A molecular phylogeny of Insecta. Arthropod System. Phyl. 64, 35–44. Klass, K.-D., 1995. Die Phylogenie der Dictyoptera. Ph.D. Thesis, Munich University. Cuvillier, Go¨ttingen, iii + 256pp. + unpaginated plates. Klass, K.-D., 1997. The external male genitalia and the phylogeny of Blattaria and Mantodea. Bonn. Zool. Monogr. 42, 1–341. Klass, K.-D., 1998. The proventriculus of the Dicondylia, with comments on evolution and phylogeny in Dictyoptera and Odonata (Insecta). Zool. Anz. 237, 15–42. Klass, K.-D., 2001. Morphological evidence on blattarian phylogeny: ‘‘phylogenetic histories and stories” (Insecta, Dictyoptera). Dtsche. Entomol. Z. 48, 223–265. Klass, K.-D., 2003. Relationships among the principal lineages of Dictyoptera inferred from morphological data. Entomol. Abh. 61, 134–137. Klass, K.-D., Meier, R., 2006. A phylogenetic analysis of Dictyoptera (Insecta) based on morphological characters. Entomol. Abh. 63, 3–50.

816

K.-D. Klass et al. / Molecular Phylogenetics and Evolution 46 (2008) 809–817

Kristensen, N.P., 1991. Phylogeny of extant hexapods. In: CSIRO (Ed.), The Insects of Australia, second ed., vol. 1. Melbourne University Press, pp. 125–140. Liechti, P.M., Bell, W.J., 1975. Brooding behavior of the Cuban burrowing cockroach Byrsotria fumigata (Blaberidae, Blattaria). Insect. Soc. 22, 35–46. Lilburn, T.G., Kim, K.S., Ostrom, N.E., Byzek, K.R., Leadbetter, J.R., Breznak, J.A., 2001. Nitrogen fixation by symbiotic and free-living spirochetes. Science 292, 2495–2498. Lo, N., Bandi, C., Watanabe, H., Nalepa, C., Beninati, T., 2003. Evidence for cocladogenesis between diverse dictyopteran lineages and their intracellular endosymbionts. Mol. Biol. Evol. 20, 907–913. Lo, N., Beninati, T., Stone, F., Walker, J., Sacchi, L., 2007a. Cockroaches that lack Blattabacterium endosymbionts: the phylogenetically divergent genus Nocticola. Biol. Lett. 3, 327–330. Lo, N., Engel, M.S., Cameron, S., Nalepa, C.A., Tokuda, G., Grimaldi, D., Kitade, O., Krishna, K., Klass, K.-D., Maekawa, K., Miura, K., Thompson, G.J., 2007b. Save Isoptera. A comment on Inward et al.. Biol. Lett. 3, 562–563. doi:10.1098/rsbl.2007.0264. Lo, N., Tokuda, G., Watanabe, H., Rose, H., Slaytor, M., Maekawa, K., Bandi, C., Noda, H., 2000. Evidence from multiple gene sequences indicates that termites evolved from wood-feeding cockroaches. Curr. Biol. 10, 801–804. Lorenc, W., 1939. Untersuchungen an Flagellaten aus dem Genus Lophomonas, I, II. Zoologica Poloniae 3, 225–250. Maekawa, K., Kon, M., Araya, K., 2005b. New species of the genus Salganea (Blattaria, Blaberidae, Panesthiinae) from Myanmar, with molecular phylogenetic analyses and notes on social structure. Entomol. Sci. 8, 121–129. Maekawa, K., Matsumoto, T., 2000. Molecular phylogeny of cockroaches (Blattaria) based on mitochondrial COII gene sequences. Syst. Entomol. 25, 511–519. Maekawa, K., Park, Y.C., Lo, N., 2005a. Phylogeny of endosymbiont bacteria harbored by the woodroach Cryptocercus spp. (Cryptocercidae: Blattaria): molecular clock evidence for a late Cretaceous-early Tertiary split of Asian and American lineages. Mol. Phyl. Evol. 36, 728–733. Maekawa, K., Terayama, M., Maryati, M., Matsumoto, T., 1999. The subsocial wood-feeding cockroach genus Salganea Sta˚l from Borneo, with description of a new species (Blaberidae: Panesthiinae). Oriental Insects 33, 233–242. Masaki, S., Shimizu, T., 1995. Variability in wing form of crickets. Res. Popul. Ecol. 37, 119–128. Matsumoto, T., 1987. Colony compositions of the subsocial woodfeeding cockroaches Salganea taiwanensis Roth and S. esakii Roth (Blattaria: Panesthiinae). In: Eder, J., Rembold, H. (Eds.), Chemistry and Biology of Social Insects. Verlag J. Peperny, Mu¨nchen, p. 394. McKittrick, F.A., 1965. A contribution to the understanding of cockroach—termite affinities. Ann. Entomol. Soc. Am. 58, 18–22. Mo¨hn, E., 1984. System und Phylogenie der Lebewesen, vol. 1. E. Schweizerbart’sche Verlagsbuchhandlung (Na¨gele & Obermiller), Stuttgart, xxxv + 834pp. Nalepa, C.A., 1984. Colony composition, protozoan transfer and some life history characteristics of the woodroach Cryptocercus punctulatus Scudder. Behav. Ecol. Sociobiol. 14, 273–279. Nalepa, C.A., 1988a. Cost of parental care in Cryptocercus punctulatus Scudder (Dictyoptera: Cryptocercidae). Behav. Ecol. Sociobiol. 23, 135–140. Nalepa, C.A., 1988b. Reproduction in the woodroach Cryptocercus punctulatus Scudder (Dictyoptera: Cryptocercidae): mating, oviposition and hatch. Ann. Entomol. Soc. Am. 81, 637–641. Nalepa, C.A., 1990. Early development of nymphs and establishment of hindgut symbiosis in Cryptocercus punctulatus (Dictyoptera: Cryptocercidae). Ann. Entomol. Soc. Am. 83, 786–789. Nalepa, C.A., 1994. Nourishment and the evolution of termite eusociality. In: Hunt, J.H., Nalepa, C.A. (Eds.), Nourishment and Evolution in Insect Societies. Westview Press, Boulder, pp. 57–104.

Nalepa, C.A., Bell, W.J., 1997. Postovulation parental investment and parental care in cockroaches. In: Choe, J.C., Crespi, B.J. (Eds.), Social Behavior in Insects and Arachnids. Cambridge University Press, Cambridge, pp. 26–51. Nalepa, C.A., Byers, G.W., Bandi, C., Sironi, M., 1997. Description of Cryptocercus clevelandi (Dictyoptera: Cryptocercidae) from the northwestern United States, molecular analysis of bacterial symbionts in its fat body, and notes on biology, distribution and biogeography. Ann. Entomol. Soc. Am. 90, 416–424. Nalepa, C.A., Li, L., Lu, W., Lazell, J., 2001a. Rediscovery of the woodeating cockroach Cryptocercus primarius (Dictyoptera: Cryptocercidae) in China, with notes on ecology and distribution. Acta Zootax. Sinica 26, 184–190. Nalepa, C.A., Bignell, D.E., Bandi, C., 2001b. Detritivory, coprophagy, and the evolution of digestive mutualisms in Dictyoptera. Insect. Soc. 48, 194–201. Noirot, C., 1985. Pathways of caste development in the lower termites. In: Watson, J.A.L., Okot-Kotber, B.M., Noirot, C. (Eds.), Caste Differentiation in Social Insects. Pergamon Press, New York, pp. 41–57. Park, Y.-C., Grandcolas, P., Choe, J.C., 2002. Colony composition, social behavior and some ecological characteristics of the Korean woodfeeding cockroach (Cryptocercus kyebangensis). Zool. Sci. 19, 1133– 1139. Parker, S.P. (Ed.), 1982. Synopsis and Classification of Living Organisms, vol. 1. McGraw-Hill, New York etc. Pellens, R., D’Haese, C.A., Belle´s, X., Piulachs, M.-D., Legendre, F., Wheeler, W.C., Grandcolas, P., 2007. The evolutionary transition from subsocial to eusocial behaviour in Dictyoptera: phylogenetic evidence for modification of the ‘‘shift-in-dependent-care” hypothesis with a new subsocial cockroach. Mol. Phyl. Evol. 43, 616–626. doi:10.1016/j.ympev.2006.12.017. Pellens, R., Grandcolas, P., Domingos de Silva-Neto, I., 2002. A new and independently evolved case of xylophagy and the presence of intestinal flagellates in the cockroach Parasphaeria boleiriana (Dictyoptera, Blaberidae, Zetoborinae) from the remnants of the Brazilian Atlantic forest. Can. J. Zool. 80, 350–359. Roach, A.M.E., Rentz, D.C.F., 1998. Blattodea. In: CSIRO (Ed.), Zoological Catalogue of Australia, vol. 23. Australian Biological Resources Study, pp. 21–162. Roisin, Y., 1990. Termite workers: a model for the study of social evolution. In: Veeresh, G.K., Mallik, B., Viraktamath, C.A. (Eds.), Social Insects and the Environment. Oxford & IBH Publishing Co., New Delhi, India, pp. 335–336. Roth, L.M., 1979. A taxonomic revision of the Panesthiinae of the World II. The genera Salganea Sta˚l, Microdina Kirby and Caeparia Sta˚l (Dictyoptera: Blattaria: Blaberidae). Aust. J. Zool. Suppl. Ser. 69, 1– 201. Roth, L.M., 1982. Introduction. In: Bell, W.J., Adiyodi, K.G. (Eds.), The American Cockroach. Chapman & Hall, London, New York, pp. 1– 14. Roth, L.M., Willis, E.R., 1960. The biotic associations of cockroaches. Smithsonian Misc. Collect. 141, 1–470. Rugg, D., 1987. Aspects of the biology of an Australian wood-feeding cockroach, Panesthia cribrata (Blattodea: Blaberidae). Master’s Thesis, Department of Plant Pathology and Agricultural Entomology. University of Sydney, Sydney, 170 pp. Sanchez, C., Hernandez, F., Rivera, P., Calderon, O., 1994. Indigenous flora in cockroaches (Dictyoptera: Blattidae and Blattellidae): a bacteriological and ultrastructural analysis. Rev. Biol. Trop. 42 (Suppl. 2), 93–96. Seelinger, G., Seelinger, U., 1983. On the social organization, alarm and fighting in the primitive cockroach Cryptocercus punctulatus. Z. Tierpsychologie 61, 315–333. Speight, M.C.D., 1989. Saproxylic invertebrates and their conservation. Council of Europe, Nature and Environment Series, No. 42, Strasbourg, 79 pp. Steinhaus, E.A., 1946. Insect Microbiology. Comstock Publishing Co., Inc., Ithaca, New York, 763 pp.

K.-D. Klass et al. / Molecular Phylogenetics and Evolution 46 (2008) 809–817 Terry, M.D., Whiting, M.F., 2005. Mantophasmatodea and phylogeny of the lower neopterous insects. Cladistics 21, 240–257. Tokuda, G., Lo, N., Watanabe, H., Arakawa, G., Matsumoto, T., Noda, H., 2004. Major alteration of the expression site of endogenous cellulases in members of an apical termite lineage. Mol. Ecol. 13, 3219– 3228. Thorne, B.L., Carpenter, J.M., 1992. Phylogeny of the Dictyoptera. Syst. Entomol. 17, 253–268. Trueman, J.W.H., Pfeil, B.E., Kelchner, S.A., Yeates, D.K., 2004. Did stick insects really regain their wings? Syst. Entomol. 29, 138–139.

817

Whiting, M.F., Bradler, S., Maxwell, T., 2003. Loss and recovery of wings in stick insects. Nature 421, 264–267. Wolcott, G.N., 1948. The insects of Puerto Rico. J. Agric. Univ. Puerto Rico 32, 1–224. Yamin, M.A., Trager, W., 1979. Cellulolytic activity of an axenicallycultivated termite flagellate, Trichomitopsis termopsidis. J. Gen. Microbiol. 13, 417–420. Zurek, L., Keddie, B.A., 1996. Contribution of the colon and colonic bacterial flora to metabolism and development of the American cockroach Periplaneta americana L. J. Insect Physiol. 42, 743–748.