For the EDS we had the herbarium records, so we could work out when data were brand new, or records for which herbarium specimens already existed.

The data show that the herbarium data reached effective zero at about 55 000 SRS. We have thus stopped getting new records for herbarium data and thus are finished - but only just.

(The slight upcurve in the graphs is because we went hell for leather to fill in the gaps during the final stages of the project - a team of four atlassers spent two weeks in every month visiting all the remote areas we could identify)

For the brand new data however, we have not yet reached zero, we are still getting in data at 1 new record for every 12-20 SRS sent in. According to extrapolations, we would have reached no new records at 80 000 SRS. In other words, we are only 75% of the way there. Three quarters of the way towards achieving our goal. At current rates of data income, that would require 20 month more of atlassing.

But there is a problem.

On the EDS scale we obtained 9356 records of proteas. The herbarium records before the atlas began covered 5903 EDS. We thus almost doubled the amount of data available at the EDS scale. However, some 1479 EDS of herbarium records were never atlassed. In other words, one quarter of herbarium records at the EDS were not found! This is a huge discrepancy!

Why?

We don’t know. To find this out we would have to do a lot more field work. They may be extinct populations, extremely isolated clumps that are too small to find easily, bad localities due to errors in recording place names or typing up herbarium labels, errors in capturing the herbarium data onto the computer, or bad identification of specimens at the herbarium, or something else. Without more field work we can never know.