I have a raft of wild willows on my place that I'm trying to identify. Now salix is a confusing species at the best of times. Keys are fine, but all too often you either have a vague division (capsules 4-7 mm long, versus capsules 3-9 mm long) or the key is on characteristics available only at a very short time of year. (Keys that depend on knowing both the shape of the anthers in April, and the shape of the capsules in June are aggravating.) I'm looking for a site where I can see multiple examples of photos of various aspects of willows to help in their ID. That is: * Leaves when young * Leaves when old * Leaves with various margins if the species shows margin variation. * Male and female catkins at various stages. * Buds at various stages. * Summer & winter twig * Bark. Ideally such a site also allows you to put in known information to sieve the site. E.g. If you fill in bud scale colour = red, you only get the ones that have been described as red. Twig color brown eliminiates the ones with other twig color. Put in the dimensions of a typical leaf, and it calculates the length/width ratio. Where possible it will ask for information that would split the remaining group the most. E.g. It may ask about leaf scars, margin. Such a website exist?
Darn. Other ideas for such a site: * If you put in your location, it would eliminate (or mark as 'out of bounds') species that didn't grow there. The range to out of bounds should be settable -- with climate change we will see ranges moving. Furthermore because of ornamental imports, range is a bit sloppy. E.g. Laurel leaf willow has escaped into the wild locally and is not native. Some users would want to identify horticultural specimens. * Put in your zone and 'zone tolerance' and it would eliminate all way out of zone selections. E.g. I'm zone 3. Some zone 4 stuff grows here. But zone 5 corkscrew willow failed miserably. * Put in neighboring recognized plants, and it would focus your search on the corresponding willows. E.g. the presence of marsh marigolds makes alpine and upland willows unlikely. The willows would make a good base project for doing this: * They are world wide. * There's a fair number of species. * They are ecologically important. * They are hard for the layman to identify.
Interesting idea, but . . . That wouldn't apply everywhere; here, Marsh Marigolds do often occur at high altitudes and could easily occur with upland willows. Variation like this around the world would make the whole idea very hard to implement.
It wouldn't be easy. The system has to be designed so that it 'learns' from it's database. The US forestry sylvics manuals list associated species for each species mentioned. There apparently are also several elaborate forest classification schemes, and I've seen lists of forest types for given species too. This is why I suggested that starting with a single moderately large, complex genus is a good place to start. The problem is small enough to be doable. Large enough that most of the gotchas that would occur in a universal database would come up. I'm using upland in the sense of upland game birds -- land well above the lowland by creeks. Upland is the counterpart of riparian. But upland does not mean alpine. I've heard green alder called upland alder -- it grows in non-riparian areas. -- compared to river alder. Golden willow here is happy as a clam on top of hills. Pussy willow and coyote willow are not. My choice of marsh marigold may have been poor. Some species are specific enough to be readily identifiable, and act as index species for an ecology. E.g. in north America if I were trying to figure out if a given larch was siberian or native tamarack, knowing that it was in a labrador tea bog is a pretty good indicator that it's tamarack, or at least no siberian.
It's been done -- sort of. Done without pictures. For north american willows (native and some introduced) see: http://aknhp.uaa.alaska.edu/willow/ The database system is described at: DEscriptive Language for TAxonomy -- Delta. http://delta-intkey.com/
