Sir

M. Soler in Correspondence (“How inbreeding affects productivity in Europe”, Nature 411, 132; 2001) quantifies endogamy, or “inbreeding”, in 14 European countries as the percentage of staff trained at the same university. His results are of great value as a basis for political decisions on university reform. He discovered a huge variation in endogamy, from 1% in Germany to 88% in Spain and 91% in Portugal. To uncover the causes, I have analysed some economic and demographic data published in the Eurostat Yearbook 2001 (Eurostat, Luxembourg, 2001; http://europa.eu.int/comm/eurostat/).

The percentage of GDP invested in research and development is significantly negatively correlated with endogamy (rs = −0.75, P = 0.0018). In terms of annual per capita investment in R&D, Portugal (67 euros; US$60.6) and Spain (116 euros) invest the least, and Sweden (936 euros) and Switzerland (935 euros) the most, with the remaining countries showing intermediate values. This indicates that endogamy is a consequence of poor investment policies.

Because there has been much discussion about Spain, I will refer to it here as an example. Spain is the second-highest country in Soler's endogamy index, but has the lowest cost per paper appearing in the Science Citation Index (SCI). This leads one to ask whether endogamy is indeed negative for the R&D system?

To answer this question, I calculated three indexes of productivity. One is simply the number of SCI papers per capita, which is negatively correlated with endogamy (rs = −0.73, P = 0.003). The second is the number of SCI papers per researcher. I obtained the number of researchers (including technicians) per country (except the United Kingdom and Germany) from the World Bank web page (http://wbln0018.worldbank.org/psd/compete.nsf/f14ea5988b0eec7f852564900068cbfd?OpenView). This second index fails to show significant correlation with endogamy (rs = −0.17, P = 0.59). My third index is the cost per paper appearing in the SCI, calculated as the quotient between the absolute amount of money invested in R&D and the number of papers appearing in the SCI. This index also failed to show significant correlation with endogamy (rs= − 0.37, P = 0.19).

Taken as a whole, these results could reflect the unequal numbers of researchers per 10,000 inhabitants, varying from 1 in Spain and Portugal to 6 in Sweden, and perhaps the different effort demanded of researchers in different countries.

In my opinion, Spanish endogamy is caused by low investment in R&D (0.9% of GDP in 1999). Spanish universities lack research technicians, so researchers have to do everything: collecting materials in the field, preparing samples for analysis, performing all laboratory techniques, and so on, as well as teaching. Spanish SCI papers are thus low-cost, but single researchers have very difficult lives, impelling us to integrate into groups for support — a process which is also encouraged by the policies of regional governments. On average, we need 10 or more years to create a research group with acceptable productivity.

Moving is very difficult for many reasons, some discussed by previous correspondents. Those who do move have to start a new lab from scratch with no equipment and without colleagues from the previous lab. The logical consequence of these factors is reluctance to move, and hence endogamy.

I agree with Soler and other correspondents that this situation needs to be corrected. My analysis shows that the best and simplest solution for Spain is significantly increased investment in R&D, which would avoid the causes of endogamy. But I am afraid that the government's proposed reform is opting for the cheapest solution, not the best one.