Researchers who make keys are not different from the rest of us; they just have learned to observe in such a way as to spot the important differences between taxa. It occurred to me that if I set my hand to making a key, that would help me to hone my observational skills in the same way.
What taxon? For this practice run, it doesn’t have to be a taxon for which no key exists. It also doesn’t have to follow phylogeny; there are what are known as “artificial” keys, which focus on identification without being concerned with taxonomic relatedness.
I decided on the genus Heniochus of Indo-Pacific butterflyfishes. This genus has enough species that a key is meaningful; enough similarities that a key is needed; and enough differences that a key is useful. Also, for my purposes, it would have to be one for which the key does not require miscroscopy; and I didn’t want to mess around with fin ray and scale counts, because those are not useful in the field. An ichthyologist with a collecting permit is one thing, but most of us oughtn’t to go around killing fish just to identify them. I didn’t have physical specimens to work with, but only photographs from various sources; but that’s okay, because a key meant for use on iNat should be designed to work with photographs.
Anyway, the key I wrote can be seen at this Google Doc:
I would like suggestions as to how to optimize it; that is, achieve reliability without excessive complexity. It is often possible to add more and more detail to a key in the quest for complete reliability, but there is a point at which it becomes overcomplex and therefore less useful. A key is not meant to be a full description of each species, but only of the distinguishing features for telling one from another.
I recommend to those of you who would like to improve your observational skills that you try your hand at creating an identification key. You don’t need to include all the species in a genus, or even limit yourself to one genus; think about what set of taxa in your area are frequently confused with each other, and create a key for separating those taxa in that locality.