have worked in coffee pollination and in some sense there are similar
challenges because the flowers are crowded in a branch, not in an
inflorescence as in your case, but push you to think in kind of similar
alternatives. Some ideas that came to my mind were:
Since the number of flowers in a raceme is too large (and surely you
don't want to count them all in each raceme of your experiment) I would
try to see if there is a good-enough correlation between total raceme length and number of flowers per raceme. A friend of mine did something similar
in to avoid a long process too - like counting flowers in your
case - and it worked. He got a correl. coef higher than 0.9 so he went
for that option. If you are lucky too - which I think is
highly possible given the related variables - you could be able to
measure total raceme length and estimate the number of flowers from that.
Based on the 'hopefully good correlation' you could get from above, you
could do some pilot pollination experiments to see if there is a bee pollination effect.
you interest behind this experiment is to provide food (the Prosopis
pods) for animals - as one of the common uses of Prosopis - I would try
to relate the pollination experiments to pod weight.
However, after weighting the pods, you can also select just some of them
to count seeds too (and why not, test if the pod weight of those
subsamples is related with number of seeds per pod). Doing that you will
get a measurement of seed set: if the relation between raceme-lenght is
good, you will be able to estimate number of flowers from that and
relate that number with the number of seeds per
I hope my comments will help you, though they are
pretty basic. I know that when one deals with plant species like
Propopis, where flowers are crowded in an inflorescence, pollination
experiments could be tricky to design.
Jaime A. FlorezPhD student
Dept. Biology & Ecology Center. Utah State Univ.
Logan, Utah. USA