Every Asphondylia gall is an ambrosia gall. A fungus lines its inner wall, and the larva feeds on that fungus as much as on the plant: the goldenrod builds the chamber, the fungus makes the food. It is the arrangement the blister-gall midge runs a few pages over, and Asteromyia is in fact the only goldenrod gall midge whose larvae have actually been shown to eat the fungus rather than the plant; for Asphondylia it is read from the lining.27

The galls give almost nothing away. A bud rosette on one goldenrod looks like a bud rosette on the next, and the flat snap galls on the leaves of tall goldenrod, rough goldenrod and Euthamia are built alike although they belong to different midges. The adults are barely more help. Telling these species apart, and working out which plant each one uses, has had to be done with DNA.27

Some of them keep two galls going. A. monacha is bivoltine: a slight, quick gall in young spring shoots on one host, and a wholly different gall in late summer on another. Others run several generations and shift as they go, from bud to leaf, or from one goldenrod to the next. Dorchin's phylogeny reads the whole group as a recent radiation, one that spread by moving between host plants and between the parts of a plant rather than by any change a lens can catch.27

Almost none of this was on record before 2015. Dorchin and colleagues described three new goldenrod species (A. pseudorosa, A. rosulata and A. silva), redescribed A. monacha and A. solidaginis, the latter's original specimens long lost so that a new name-bearing one had to be set, and struck off A. johnsoni, named in 1908 from an unknown gall on an unknown goldenrod, as a name that means nothing. Seven goldenrod species in all, most of them new to science.27

The galls are crowded, and what comes out is often not the Asphondylia. A second midge, Youngomyia podophyllae, lives in them as an inquiline; a wasp, Galeopsomyia haemon, works them as a parasitoid; between them they are among the commonest things to emerge.27

So the biggest genus in this catalogue is the one it knows least. Most of its forms are undescribed bud rosettes, separated, where they are separated at all, by the goldenrod they were found on. It is the plainest measure in the guide of how much goldenrod-gall work is still to do.