Field of Science

Stars in the Pasture

Flowers of Stellaria graminea, photographed by Jiří Bohdal.

Last week, I described a situation where low morphological diversity disguised high chromosomal and species diversity. So it is fitting that, this week, I should be focusing on an almost opposite situation: high morphological and chromosomal diversity, but low species diversity.

Stellaria is a genus of weedy herbaceous plants native primarily to Northern Hemisphere temperate regions, though some species have been spread around the world by humans. Members of the genus are mostly characterised by deeply divided petals, so that at a glance a Stellaria flower may look to have twice as many petals as it really does. (It is notable in this regard, however, that Chinnappa & Morton (1984) recorded raising a single seedling of Stellaria longipes during breeding tests that produced flowers with entire rather than divided petals, a feature generally characteristic of other related genera.) Up to about 200 species have been assigned to the genus (Bittrich 1993) but, considering what has been established with the American S. longipes complex (read on), I would not be at all surprised if that number is somewhat inflated. A review of Chinese Stellaria by Wu (1991), for instance, reduced a third of the species treated to synonyms.

Stellaria longipes, photographed by Michael Shepard.

The issue, as already alluded, is that some of the widespread Stellaria species are extremely variable. A case study of variability in Stellaria has been conducted with members of the Holarctic S. longipes complex (Macdonald et al. 1988). Four taxa are currently recognised within this group: Stellaria longipes ssp. longipes and S. longifolia both have circumpolar distributions in North America and Eurasia, but S. longipes ssp. arenicola and S. porsildii are each known from only small ranges in North America (Chinnappa 1992). Stellaria longifolia and S. porsildii are diploid species with a chromosome number of 26; S. longipes is a polyploid species, possibly derived from the hybridisation of the two diploids.

Specimens attributed to Stellaria longipes show a huge variation in morphological characters: stem pubescence, growth habit, floral and fruit characteristics, etc. They also show a wide variation in chromosome numbers: individuals have been recorded with as few as 51 or as many as 107 chromosomes. You may be thinking that such variation must indicate that a number of species have been conflated under the name of 'S. longipes', and, indeed, a number of names have been based on variant forms of this complex. However, breeding tests conducted by Chinnapa & Morton (1984) failed to find any reproductive isolation between different morphs: all were fully interfertile. What is more, different morphs are not found in different ranges, supporting the cultivation results as a valid representation of what is happening in the wild. Despite their variability in morphology, the representatives of S. longipes form a single, rampantly outbreeding gene pool. Even the apparent question of chromosome variability becomes less significant when one considers that the majority of individuals have chromosome numbers of 52, 78 or 104: all multiples of 26. If you compare this to what we find in Rhogeessa, you can see that it is not chromosome number alone that is the controlling factor for interfertility. The important thing is the ability or otherwise of the chromosomes to form associations during meiosis allowing their appropriate segregation to form viable gametes.

A slightly different form of Stellaria longipes, from here.

The only segregate of Stellaria longipes that appears to be truly distinct is the form known as S. longipes ssp. arenicola. This is a restricted range form known only from an area of sand dunes south of Lake Athabasca at the northern end of the border between Alberta and Saskatchewan (true S. longipes ssp. longipes is also found in the same area). Distinction appears to have arisen between S. l. ssp. longipes and S. l. ssp. arenicola as a result of a change in fertilisation regime for the latter. Stellaria species (including S. longipes) are mostly protandrous: the male parts of the flower mature before the female parts are ready to receive pollen. This means that a flower generally cannot fertilise itself, and individuals are generally outcrossing. Stellaria longipes ssp. arenicola has reversed this pattern: protandry is reduced, and individuals are more likely to be fertilised by themselves than by others. The nature of their sand dune habitat also means that plants are more likely to germinate successfully alongside another individual than in an isolated position. This results in a clumped and patchy distribution of individuals that also promotes fertilisation within a limited pool of individuals, as fertilising insects forage within a single patch rather than travelling between patches. Genetic analysis indicates that a limited amount of gene flow probably does still occur between sympatric S. longipes ssp. arenicola and S. l. ssp. longipes (Purdy et al. 1994); nevertheless, this outcrossing is on the order of the occasional dalliance rather than the shameless harlotry of their conspecifics elsewhere.


Bittrich, V. 1993. Caryophyllaceae. In: Kubitzki, K., J. G. Rohwer & V. Bittrich (eds) The Families and Genera of Flowering Plants vol. 2. Flowering Plants: Dicotyledons: Magnoliid, hamamelid and caryophyllid families pp. 206-236. Springer.

Chinnappa, C. C. 1992. Stellaria porsildii, sp. nov., a new member of the S. longipes complex (Caryophyllaceae). Systematic Botany 17 (1): 29-32.

Chinnappa, C. C., & J. K. Morton. 1984. Studies on the Stellaria longipes complex (Caryophyllaceae)-biosystematics. Systematic Botany 9 (1): 60-73.

Macdonald, S. E., C. C. Chinnappa & D. M. Reid. 1988. Evolution of phenotypic plasticity in the Stellaria longipes complex: comparisons among cytotypes and habitats. Evolution 42 (5): 1036-1046.

Purdy, B. G., R. J. Bayer & S. E. Macdonald. 1994. Genetic variation, breeding system evolution, and conservation of the narrow sand dune endemic Stellaria arenicola and the widespread S. longipes (Caryophyllaceae). American Journal of Botany 81 (7): 904-911.

Wu Z.-Y. 1991. Some problems on the taxonomy of Chinese Stellaria. Acta Botanica Yunnanica 13 (4): 351-368.

From Three to Two

(I've been waiting three and a half years to use Neil's icon.)

The mysterious anabaritids of the Lower Cambrian have been referred to on this site before. In the earlier, somewhat brief post, I referred to their triradial structure and uncertain, though probably coelenterate-grade, relationships. In the time since that post appeared, the anabaritids have been the subject of a review by Kouchinsky et al. (2009) that brought together a lot of the previously scattered information on these animals.

The image of an anabaritid in the previous post showed the best known species, Anabarites trisulcatus. However, this was not the only species in the group. The image just above, from Kouchinsky et al. (2009), shows another species, Anabarites biplicatus, recorded from the Siberian Platform. This species differs from A. trisulcatus in that it started out life triradial (albeit with the internal dividing ridges between the lobes only weak), but as it grew it lost its triradiality and became more bilateral (cross-section below from Kouchinsky et al.):

Some of my readers may remember that Cambrian problematica were something of a cause célèbre during the mid-90s when a lot of journals and magazines ran features on them (probably inspired to a certain degree by Stephen Jay Gould's somewhat dreadful book Wonderful Life). In the more academic corners of this pageant, the triradiality of anabaritids (as well as some other early animals such as Tribrachidium) garnered them a certain degree of attention. It was suggested by some that they might represent a unique animal lineage that was eventually superseded by our own bilateral dynasty. However, the changing symmetry of Anabarites biplicatus serves as a reminder that we should not be too hasty to assign great significance to such features. Indeed, in the modern fauna, nematodes are partially triradial (they have a triradial head structure, with one upper and two lower lips around the mouth). Though the affinities of anabaritids are somewhat debatable, the most popular scenario is that they are related to cnidarians: their tubes bear a certain resemblance to the polyps of some medusozoans. Cnidarians also exhibit a wide variety of symmetries, such as the tetraradial organisation of scyphozoans and the hexaradial organisation of hexacorals. Without preserved soft tissue to inform us what the organisation of the inhabitant animal may have been, it is difficult to say just how much weight the triradiality of the anabaritid tube should be given.


Kouchinsky, A., S. Bengtson, W. Feng, R. Kutygin & A. Val'kov. 2009. The Lower Cambrian fossil anabaritids: affinities, occurrences and systematics. Journal of Systematic Palaeontology 7 (3): 241-298.

Little Yellow Bats

Specimen of an unidentified Rhogeessa, photographed by Kate Comis.

The Neotropical members of the genus Rhogeessa go by the incredibly imaginative vernacular name of 'little yellow bats'. On the face of it, this would seem to sum up the salient features of these animals pretty succinctly: they're small, they're yellow, and they're bats. Eleven species of Rhogeessa were recognised by Baird et al. (2008a), found from northern Mexico to southern Brazil. The species of Rhogeessa are rather difficult to distinguish from each other; in particular, the six species that referred by Baird et al. (2008a) to the 'Rhogeessa tumida complex' are all but indistinguishable. Nevertheless, recent authors have regarded them as good species, and it is because of the reasons why this is that the genus has attracted the most interest.

The only really reliable way to distinguish the species of the R. tumida complex is to take a look at their chromosomes. Despite their external similarity, the species have different chromosome numbers and arrangements from each other. This forms an interesting contrast to other bat genera, which may have more morphological variability but little chromosome variation. Comparison between Rhogeessa chromosomes has lead to the suggestion that the various species may have diverged as a result of a process called Robertsonian translocation.

Specimen of the Yucatan yellow bat Rhogeessa aeneus photographed by Alex Borisenko.

To explain Robertsonian translocation, I have to indulge in a bit of background terminology of chromosomes (skip this if you know all this stuff). Think of the classic picture of an X-shaped chromosome (this is actually a doubled chromosome that develops during cell division: two copies, called chromatids, have been produced of the chromosome that will be separated when the cell divides). The point where the two chromatids are joined is a region of the chromosome called the centromere: it provides an attachment point for the spindle fibres that will draw each individual chromatid apart. The centromere is not always positioned at the midpoint of the chromosome: those chromosomes in which it is are called metacentric, while other acrocentric chromosomes have the centromere close to one end so the conjoined chromatids look closer to V-shaped than X-shaped.

Translocation is a process where a piece of genetic material breaks off one chromosome and becomes attached to another. Robertsonian translocation is a particular type of translocation where two acrocentric chromosomes, by breaking at the centromeres, effectively become fused to form a single metacentric chromosome (as shown in the diagram below from here):

Robertsonian translocation has been observed in many species; it even happens occasionally in humans. Where it becomes interesting for evolutionary studies is that the resulting metacentric chromosome continues to function in the same manner as the original acrocentric chromosomes, with little or no negative effects (the short bit from each acrocentric chromosome that is lost rarely contains any functioning genes). The individual carrying the fused chromosome even remains fertile, because when meiosis occurs in any individual with both the fused and unfused chromosomes, the two unfused chromosomes will each line up with their matching arm on the fused chromosome. However, imagine a situation where one individual in a population experiences a Robertsonian translocation between two chromosomes (call them 1 and 2), but another individual has a translocation between one of those chromosomes and another chromosome (say, 2 and 3). The individual that carries chromosomes 1-2 and 3 will produce fertile offspring if mating with an unfused individual, as will that carrying 1 and 2-3. However, if the 1-2 individual mates with the 2-3, their offspring will carry both fused chromosomes. Because these chromosomes and the unfused 1 and 3 cannot easily match up in a way that allows them to be separated effectively during meiosis, the hybrid offspring will have significantly reduced fertility. This has been practically shown to be the case between Robertsonian races of mice (Capanna et al. 1977). If the two fused chromosomes each become more predominant in a population than the original unfused chromosomes (either by drift or hitchhiking), then gene flow will be slowed or stopped between individuals carrying one or the other. Hey presto, speciation!

Speciation as a result of Robertsonian translocation also provides a counter-example to those who, when objecting to the taxonomic recognition of 'cryptic' species, raise the Biological Species Concept to defend their viewpoint. Contrary to popular assumption, there is no essential correlation between speciation and morphological divergence. Even under the Biological Species Concept, two populations may be good species (i.e. non-interfertile) and yet morphologically indistinguishable.

Allen's yellow bat Rhogeessa alleni, photographed by Merlin Tuttle. Rhogeessa alleni is the most morphologically distinct species of Rhogeessa, and has been included in a separate genus or subgenus Baeodon. Baird et al. (2008a), on the basis of molecular phylogeny, recognised the subgenus Baeodon but also including R. alleni's sister species, R. gracilis.

Anyway, I have a vague memory that somewhere along the line I was talking about bats. The known karyotype numbers for Rhogeessa vary from 30 (in three species: R. alleni, R. gracilis and R. io) to 52 (in a specimen from Suriname that Baker et al. 1985 assigned to R. tumida but which almost certainly represents an undescribed species). The phylogeny for the genus that was recovered by Baird et al. (2008a, b) could be consistent with both Robertsonian fusions and fissions taking place during the genus' history. In an attempt to test whether the different karyotypes truly function as isolating mechanisms (and hence whether the chromosomal 'species' are actually species), Baird et al. (2008b) could only find genetic indicators of possible recent hybridisation between the two apparently least divergent species, R. tumida (34 chromosomes) and R. aeneus (32 chromosomes); all other species maintained reciprocal monophyly in each of the three gene types (mitochondrial, Y-chromosome and somatic chromosome) tested. Baker et al. (1985) referred to another 32-chromosome karyotype ('32N') that differed from R. aeneus (assuming, on the basis of geography, that R. aeneus corresponds to Baker et al.'s '32B') in terms of exactly which chromosomes had been fused and so would be reproductively incompatible with R. aeneus. However, this 32N form appears to be assigned by Roots & Baker (2007) to R. io, otherwise with 30 chromosomes. As 30-chromosome R. io and the 32N form differ only in a single pair fusion in the former (Baker et al. 1985), they would probably remain interfertile by the principles described earlier. In contrast, despite their apparent difference in chromosome number of only two, R. tumida and R. aeneus actually differ in five chromosome fusions (three on one side, two on the other), meaning their interfertility should be considerably lower.

And if you've gotten this far and you're still not sick of bats, Darren Naish covered Rhogeessa and its relatives as part of his mammoth series on vesper bats earlier this year.


Baird, A. B., D. M. Hillis, J. C. Patton & J. W. Bickham. 2008a. Evolutionary history of the genus Rhogeessa (Chiroptera: Vespertilionidae) as revealed by mitochondrial gene sequences. Journal of Mammalogy 89 (3): 744-754.

Baird, A. B., D. M. Hillis, J. C. Patton & J. W. Bickham. 2008b. Speciation by monobrachial centric fusions: A test of the model using nuclear DNA sequences from the bat genus Rhogeessa. Molecular Phylogenetics and Evolution 50 (2): 256-267.

Baker, R. J., J. W. Bickham & M. L. Arnold. 1985. Chromosomal evolution in Rhogeessa (Chiroptera: Vespertilionidae): possible speciation by centric fusions. Evolution 39 (2): 233-243.

Capanna, E., M. V. Civitelli & M. Cristaldi. 1977. Chromosomal rearrangement, reproductive isolation and speciation in mammals. The case of Mus musculus. Bolletino di Zoologia 44 (3): 213-246.

Roots, E. H., & R. J. Baker. 2007. Rhogeessa parvula. Mammalian Species 804: 1-4.

The Beak-Shells' Legacy

Specimen of Conocardium japonicum, from here.

The Rostroconchia were a group of molluscs that lived during the Palaeozoic, being definitely found from the Late Cambrian to the latest Permian. The name means, roughly, 'beak-shell', and refers to the shape of the shell. Similar to bivalves (with which rostroconchs were classified prior to the 1970s), rostroconchs had a shell divided into left- and right-hand valves, which in many species were elongate in one or both directions forewards and backwards. The shells of different species gaped to varying degrees at the posterior and anterior ends. Rostroconchs differed from bivalves, however, in that the original larval shell was not divided. Instead, the shell of a rostroconch was initially cap-shaped and entire. As the animal developed past the larval stage, the shell developed lateral lobes that would eventually become the two valves. As a result, rostroconchs did not have a toothed hinge connecting the valves like that of bivalves. Nevertheless, in many species, the pressure of the valves growing outwards eventually caused the larval shell to break through (Pojeta & Runnegar 1976). Like many bivalves, rostroconchs would have been infaunal, living buried in the sediment; however, the absence of a toothed hinge or the adductor muscles that connect the valves in bivalves means that the shell of rostroconchs would have been less flexible.

The Ordovician scaphopod Rhytiodentalium (topmost image) and the scaphopod-like rostroconch Pinnocaris (lower two images), from here.

The largest work to date on rostroconchs was that of Pojeta & Runnegar (1976), who suggested that rostroconchs included the ancestors of two living classes of molluscs, the bivalves and the scaphopods. Scaphopods, tusk-shells, are a group of tubular molluscs with an opening at each end of the tube that also live buried in sediment. According to Pojeta & Runnegar's suggestion, the bivalves evolved through the evolution of two calcification centres in the larval stage. The scaphopods would have evolved through the fusion of the valves along the ventral margin (as embryological studies have shown the scaphopod shell does develop), together with the restriction of growth to the anterior direction only.

However, more recent studies of molluscan phylogeny (Wilson et al. 2010) have mostly indicated that bivalves and scaphopods are not closely related relative to other molluscs. Bivalves are probably a relatively basally derived group, while scaphopods are closer to cephalopods. The upshot of this for rostroconchs is that they may be related to bivalves or scaphopods, but not to both. Of the two, it seems more likely that rostroconchs are related to scaphopods (or, as some authors have put it more bluntly, scaphopods are living rostroconchs). As well as the anatomical arguments that have been made in favour of such a relationships, a scaphopod connection has the advantage over a bivalve one of stratigraphy. The oldest definite rostroconchs, as previously noted, are known from the Late Cambrian. However, the oldest definite bivalves come from the Early Cambrian—that is, some 40 or 50 million years earlier than the first known rostroconchs. Scaphopods, on the other hand, do not appear until the Ordovician or Devonian (depending on whether the earlier forms are accepted as scaphopods), well after the known appearance of their supposed ancestors. Pojeta & Runnegar (1976) did not recognise a stratigraphic conflict in deriving bivalves from rostroconchs as they had identified the Early Cambrian Heraultipegma as a rostroconch. This identification, however, was disputed by MacKinnon (1985), who held that the characters cited by Pojeta & Runnegar in support of their assignation had been misinterpreted and were not truly present in Heraultipegma.

The conocardioid rostroconch Arceodomus longirostris, from here.

Pojeta & Runnegar (1976) recognised three orders of rostroconch: the Ribeirioida, Ischyrinioida and Conocardioida. This division has been followed by all subsequent authors, though it should be noted that Pojeta & Runnegar recognised the Ribeirioida as explicitly paraphyletic with regard to the other orders. Pojeta & Runnegar (1976) regarded the scaphopods as derived from ribeirioids (specifically related to the somewhat scaphopod-like Pinnocaris), but Peel (2004) stated that protoconch characters indicated a derivation from conocardioids. The ribeirioids and ischyrinioids both became extinct at the end of the Ordovician, leaving only the conocardioids and their descendent scaphopods until the former also became extinct at the end of the Palaeozoic.


MacKinnon, D. I. 1985. New Zealand late Middle Cambrian molluscs and the origin of Rostroconchia and Bivalvia. Alcheringa 9 (1): 65-81.

Peel, J. S. 2004. Pinnocaris and the origin of scaphopods. Acta Palaeontologica Polonica 49 (4): 543-550.

Pojeta, J., Jr & B. Runnegar. 1976. The paleontology of rostroconch molluscs and the early history of the phylum Mollusca. U.S. Geological Survey Professional Paper 968: 1–88.

Wilson, N. G., G. W. Rouse & G. Giribet. 2010. Assessing the molluscan hypothesis Serialia (Monoplacophora + Polyplacophora) using novel molecular data. Molecular Phylogenetics and Evolution 54 (1): 187-193.

The Red-lined Wings of South America

So far as I've been able to find, what you see above is the only published illustration of the South American lacewing Belonopteryx arteriosa, from its original description by Gerstaecker (1863) (or very nearly from there: somewhat confusingly, the plate illustrating this animal was published an issue earlier than the article describing it). It is something of a pity that the drawing is not in colour, as Gerstaecker's description indicates a quite handsome animal: about 16 mm long with a ca 20 mm wingspan* with a mostly orange head, golden yellow body and blood red veins on the wings. The first segment of the antenna was yellow, the second darker, and the remainder black. Along the medial and secondary radial veins, the red colour of the veins extended to part of the cells on either side, producing two longitudinal red stripes on each wing.

*Gerstaecker gives the body length as 8 lines, with the front wings 9.5 lines long. A 'line' is a unit of measurement used by a number of 18th and 19th century biologists. The exact length of a line seems to have varied somewhat between countries (see this page for explanations), though it seems to have generally been a little more than 2 mm. Linnaeus apparently defined a line in the introduction to Philosophia Botanica as the length of a lunule (the white half-moon at the base of a fingernail) on any finger other than the thumb.

Despite being the type of the tribe Belonopterygini, Belonopteryx arteriosa seems to have received little attention since its description. Gerstaecker (1863) held only a single female specimen, reporting its collection locality (with a precision not uncommon for his time) as "Brazil". Freitas & Penny (2001) indicated that this species was known only from three specimens from Argentina (whether this indicates that Gerstaecker's locality was mistaken, or whether this is in addition to the original location, I couldn't say). The larva of B. arteriosa has never been identified, though larvae of related genera are associated with ant nests.


Freitas, S. de, & N. D. Penny. 2001. The green lacewings (Neuroptera: Chrysopidae) of Brazilian agro-ecosystems. Proceedings of the California Academy of Sciences 52 (19): 245-395.

Gerstaecker, A. 1863. Ueber einige neue Planipennien aus den familien der Hemerobiiden und Panorpiden. Entomologische Zeitung 24 (4-6): 168-188 (plate in issue 1-3).