My previous post was part of an exchange with Chris Buddle on whether taxonomists should describe new species without knowing their natural history. When many of the specimens upon which we base species descriptions are already long dead by the time we recognize them in a collection as something new, it’s sometimes difficult to say much about their habits and interactions. Of course, this lack of knowledge about a particular species can be offset by the predictive power of phylogeny.
One of the great revolutions in taxonomy was phylogenetic systematics, as articulated by Willi Hennig (why some taxonomists in 2013 still do not embrace the principles of phylogenetic systematics is beyond me, but that’s a rant for another post). One of the most significant implications of a phylogenetic approach is that if our classifications reflect the evolutionary history of a group, we can make predictions about unknown traits of species based on known traits of closely related species. So, if I describe a new species that fits into a genus in which all other known species are predators of snails, for example, it’s a good bet that my new species probably eats snails. But that doesn’t mean we’re always right. Most carabid beetles are predators. But not all. Most spiders are predators. But not all.
I think a lot about the limits of predictability, and the pitfalls of predicting incorrectly, in the context of the insects I know best — the fly family Chloropidae (also known as frit flies, grass flies, or eye gnats). Predictions about the natural history of poorly known species of chloropids are difficult for three reasons: 1) we only know the habits of a very small fraction of the described species; 2) many genera of chloropids are probably not monophyletic, or “natural groups”. In other words, they may include a set of species that are not each other’s closest relatives and thus don’t reflect shared history; and 3) chloropids are one of the most ecologically diverse families of insects on the planet.
A lot of reference books, websites and other resources state that the larvae of most chloropids are phytophagous, feeding on living plant tissues. That statement is almost certainly wrong. Here’s why:
In most groups of insects, the species we know best are those that have an impact on human life. In Chloropidae, the larvae of a few species are pests of cereal crops. And the closely related species in genera such as Oscinella (the group that includes the actual “frit fly”), Meromyza and Chlorops are probably phytophagous too. And it’s tempting to generalize this habit across other genera in the family. That’s where problems arise because most chloropids aren’t in these genera. The family is divided into three subfamilies; let’s look at the known habits of each.
The subfamily Siphonellopsinae is dominated by the large, mostly tropical, genus Apotropina. The larvae of the few species that are known have been reared in association with nests of social or solitary Hymenoptera, where they are apparently scavengers, or have been reared from rotting plant material (scavengers or bacterial grazers). Apotropina is a big genus with lots of described species, but there are many more undescribed species. Phytophagous? Probably not.
A lot of the plant-feeding species fit into the subfamily Chloropinae. Chlorops and Meromyza are both in this subfamily, and multiple species have been reared from grasses and sedges. Several other genera in this subfamily also include species that feed in living plants. But Thaumatomyia is a chloropine too, and the few known larvae are predators of root aphids. Another chloropine genus, Pemphigonotus, includes at least one species with a fondness for rotting crabs on ocean beaches. For the subfamily overall: Mostly phytophagous (probably), but definitely not all.
I’ve saved the biggest mess for last — subfamily Oscinellinae. The most genera and species, and the greatest range of habits. Some genera (Oscinella, Lipara, Dicraeus and others) are phytophagous. Probably. Some other genera are often assumed to be phytophagous, and you can find references to that in the published literature. But groups such as Rhopalopterum and Anatrichus are apparently secondary invaders that scavenge (or may be predators) in feeding galleries made by moths or other flies. Other oscinellines are predators of spider egg sacs, or grasshopper or mantid egg masses. Some are subcutaneous parasites of frogs in Australasia. Some kleptoparasitic species steal meals from spiders and robber flies and assassin bugs. Some feed on frass and debris in bark beetle galleries in trees. Some are scavengers in bee nests. Many are scavengers in decaying plants. Some feed on fungi. A few have a fondness for carrion. For the vast majority of species, and even genera, we simply do not know their natural history. We do know, though, that if there’s a dominant mode of life, it’s probably saprophagous, feeding in decaying organic material. Mostly phytophagous? Probably not.
There are two main obstacles to sorting out this tremendous ecological diversity in chloropid flies. The first is a taxonomic challenge — the phylogenetic relationships in this family are poorly resolved, which makes it almost impossible to construct a classification that reflects the evolutionary history (and therefore the ecological history) of the group. These flies are also so challenging to sort out morphologically that we have little hope of resolving their higher-level relationships without also incorporating DNA sequence data.
The second challenge is grounded in ecology and natural history. We need more basic observations on what these flies actually do. The habits of even our local species are still so poorly known that we could make great progress simply spending time in the field collecting likely food sources and rearing the flies. My former grad student Fred Beaulieu, now a mite specialist at the Canadian National Collection, reared several local species of chloropids and other flies from grasses, sedges and cattails, helping to document their natural history just a little bit more fully (see Beaulieu & Wheeler 2002 in Publications).
We often tend to think about research projects in terms of posing big questions and testing hypotheses or predictions and accumulating lots of data. But given how little we know about the natural history of many arthropods, we can also make significant advances in knowledge armed with a notebook, some empty pill bottles and a sunny afternoon.