Identification of orchids in fruit

Orchids flower in most cases for only 2 to 3 weeks, and most of them can be identified with certainty, though some experience is required for three genera, the fragrant orchids, the marsh orchids and the helleborines.   Hybrids are a further complication.  

Before flowering reliance on the leaf shape and arrangement can work; Poland and Clement's 'Vegetative Guide to the British Flora' is the best guide and Sean Cole and Mike Waller are working on a photographic guide to be published next year.   But post flowering, what then?  

Over the last few years, but particularly 2017 and 2018, I have been trying to develop a key for orchids in fruit, at a point where they have seed pods with viable seeds, and for most species leaves have withered away.   I have examined 30 species (out of a possible 54) and a further species where I only had a seed pods.  Of the remaining 23 species, 9 are protected and it is illegal to remove any part.  In any case they are pretty much site specific anyway as are another 5.  One has not been seen for years, 3 because of their diminutive size will be impossible to locate in fruit, so that I am left with 5 to look for in 2019, and one more needs further data.

I have measured from photos and direct measurement the angle of the fruits to the stem, the length and width of seed pods.   Using a microscope (x40 and x100) I have measured the length and width of seeds (actually the outer casing, the testa, provided that it contains an embryo.   Habitat, and extant leaves are also useful.  

Seed pod measures

Angle to stem

Seed measures

It does seem possible to discriminate, especially where for example the length / width ratio of seed pods is used rather than absolutes, and provided a reasonably large random sample has been used to determine parameters (5 seems to be enough).  

These are a couple of examples of filtering the results.

I am planning to have a poster on this at the upcoming BSBI meeting in November at Edge Hill.