Phew... it has been a long ride (>560 days) but my data processing is finally complete!
During my 3 month field season in 2016, I mapped 100 vegetation plots and collected seed from as many of the plants in those plots as I could grab (nearly 12,000 samples). I then brought those samples back to the lab and counted the seeds from each of those samples (with lots of help from others). We had to count most samples by hand because of the level of detail of the seed count - seed per pod/head - negated the usage of digital methods (e.g. taking a picture and using imageJ would probably take just as long or longer given the few number of seeds per sample for most samples). For some samples with really tiny seeds, we used a mass-count method, weighing a set number of seeds to estimate the number of seeds for that plant. In other instances, such as for grasses, we counted bracts of the culm/panicle and multiplied it by the number of seeds held in the bract.
The final seed tally is 864,748 from 12,139 samples.
Then I went through the maps to include any information for individual plants that I wrote on the map and did not collect a sample (e.g. a note of '0 seeds' on the map in the field instead of making an empty envelope). While mapping I recorded information on number of flowers/seed heads not collected (due to time constraints) by making a note next to the point, or grouping individuals into a contour on the map, or by jotting notes in the map legend. These all are varying levels of seed count quality (with the best being actually counted and the worse being a note in the legend) but this also provides more potential information for future analyses instead of just 'NA'.
Using these map notes, we can add 4,824 individuals to the seed count dataset (considering the point and contour notes).
While counting, we also made note of any seed pods that were collected to early or after dispersal (resulting in no seeds) as well as seed pods that were open ( indicating the potential for missing seed). We also had counts of number of flowers and buds still remaining on the sample. Using this data, we can estimate the number of seeds per head/pod/stem and calculate a 'corrected' seed count to use in my analyses.
So an updated count is 941,674 seeds from 13,551 individual plants (corrected for lost seed and lost potential seed - e.g. a flower still on the sample). Compared to a total of 28,579 individuals mapped across 100 plots (max=965 plants in one 50 X 50 cm plot, min=53 plants/plot, median=231 plants/plot), that is about a 47% fecundity data sampling success across the plots - not too shabby...
Now onto my analyses...