The dismal science and dismal science writing

I was a bit flippant a few days ago about the costs of genebanks. And I felt guilty enough about it, especially after Jeremy’s recent piece at Vaviblog on the value of germplasm collections, to look into it a bit more.

It all started with an article in IITA’s R4D Review which looked at the costs of conserving the cowpea collection there. The bit I had trouble with was this:

Using 2008 as a reference year, US$358,143 and $28,217 was spent annually on the conservation and management of cowpea and wild Vigna. The capital cost took the major share of the costs, followed by quasi-fixed costs for scientific staff, nontechnical labor, and nonlabor supplies and consumables. Each accession cost about $72 for cowpea and only about half of that for wild Vigna.

Now, if you know that there are something like 15,000 cowpeas in IITA’s collection, and multiply that by $72, you can very quickly see that you don’t get $386,000, and you might just start to feel justified in losing confidence in the whole exercise.

So what’s going on? Well, what’s going on, when you look into the numbers, ((Which you can do here, a document you can download from the Crop Genebank Knowledge Base.)) is that the “per accession cost” of $72 was calculated by adding up the individual per accession costs incurred for a set of about a dozen different genebank activities (from acquisition to information management to distribution), and the numbers of accessions involved in each of these was quite different, ranging from 0 for acquisition to 15,000 for long-term storage. So, for example, 475 accessions were distributed at a per accession cost of $22; 2,360 germination tested at $6 a pop etc. Add up all of these per accession costs, as incurred in 2008, and you get $72, for a total of about $386,000.

So, I’m not sure if that $72 really has much of a meaning, but at least now you know how it was calculated. Which maybe the writer of the original piece should have explained.