A Persistent Problem: We Lack a Unified Lexicon for Hymenoptera Anatomy

From HAO Wiki

The extraordinary diversity of Hymenoptera, both in terms of numbers of taxa and morphotypes, combined with the vast array of people and labs focusing on projects that span ecology, genomics, behavior, and systematics, has yielded countless partially overlapping lexicons to describe Hymenoptera anatomy. While we all share common names for broadly understood structures—such as head, wing, leg, mandible—each subcommunity of hymenopterists has created its own terms that may or may not apply to shared structures across the phylogeny of this order. For example the 'face' in bees (defined as the part of the head ventral to the medial ocellus and dorsal to the ventral margin of the head) is not homologous with the 'face' in some groups of parasitic Hymenoptera (the part of the head ventral to the antennal foramen and dorsal to the ventral margin of the head)—depending, of course, on which researcher is referring to this structure and which of the hundreds of available publications one refers to as his/her source for morphological terminology.

The disparate terminologies result in a vast sea of inconsistent (but largely recoverable) statements about hymenopterans and prohibit future efforts to explore biological phenomena across the order, including:

• Comparisons of gene expression patterns—mining genome annotations to understand the roles that genes play in morphology and behavior
• Comparative morphology and phylogenetics—increasingly important as we try to include the vast amount of information from the fossil record and include rare taxa for which we cannot extract DNA
• Descriptive taxonomy—the corpus of descriptive taxonomy (each description is essentially a block of Entity-Quality statements, similar the example at the bottom of page 1) cannot be efficient queried for information to inform us on characters and species diagnosis

How hymenopterists use language

To more explicitly understand language use in Hymenoptera taxonomy we examined five papers each for five major hymenopteran lineages that we suspected of having lexicons with the least amount of overlap: Formicidae (ants), other Aculeata (including bees and social wasps), Chalcidoidea (including the model organism Nasonia), other Proctotrupomorpha (includes the subject of the Platygastroidea PBI), and Ichneumonoidea. No individual person authored more than one paper in this set. The data clearly shed light on what was until now only a set of anecdotal observations: of the 296 anatomical terms collected from these 25 papers only 28 terms (9.5%) were used for all lineages; 155 terms (52%) were specific to a single lineage (e.g., 'anellus' was used only in papers that described Chalcidoidea).

A finer scale examination of these data indicates that the language disparity is also a taxonomist-specific problem (i.e., each domain expert has his/her own vocabulary). Only 3-5 anatomical terms (2-5%!) within a lineage overlapped across all five papers describing species in that lineage; 45-86 terms (45-62%) were specific to a particular author of species in that lineage. Some of this disparity can be attributed to the fact that taxonomists occasionally describe different structures for each taxon (e.g., the ant species descriptions might not include many statements about wing structures, whereas for wasps those parts might be important for diagnosis). We are confident, however, in stating that the tens of thousands of comparative morphology and species descriptions for Hymenoptera (authored by thousands of scientists) almost never use a common language. In fact there is no single, comprehensive resource for anatomical terminology, and authors often cobble together a series of disparate publications to use as their collective "source." Compounding this problem are the frequent, almost omnipresent, statements about the unique ways in which a particular term is used in each publication ("anellus here is defined as...").

Why it matters

As mentioned, the term 'anellus' was used only in the papers that describe chalcidoid wasps; in these cases the 'anellus' referred to a modified (smaller) subdivision of the antenna. Many other hymenopterans have these structures, but if one queried the available corpus of species descriptions only the Chalcidoidea would show up as having these subdivisions. Interestingly, some authors also use the term 'anellus' to refer to a structure in the male genitalia complex. Meaningful queries of those descriptions are therefore impossible—one would receive a series of statements about presence/appearance of antennal structures for some organisms and male genitalia for others.

Antennae are the frequent subjects of gene expression experiments, as they are often the primary structure involved in mate and host finding. Genitalic complexes, likewise, avail themselves as objects that provide critical knowledge of reproductive processes and are the primary focus for several labs that study the hymenopteran model Nasonia. One can imagine that studies with annotations that focus on gene expression in the 'anellus' would subsequently lead to confusion.