Diatom-salinity transfer functions for interpretation of palaeosalinit
y and palaeoclimate change have been developed successfully for parts
of North America and North and East Africa, but there is a need for da
ta-sets in other saline lake regions of the world. A data-set of 74 mo
dem diatom samples and associated water chemistry data is described fr
om Spain. The influence of conductivity and other environmental variab
les on diatom distribution is explored using canonical correspondence
analysis (CCA) and partial CCAs. A transfer function is derived for co
nductivity (70 samples) whose apparent predictive ability is high (app
arent r(2) = 0.91). Performance under jackknifing is poor due to the h
eterogeneous nature of the data-set and poor coverage of the freshwate
r end of the salinity gradient. There is a lack of suitable low-salini
ty sites in Spain, and the accuracy of estimated salinity optima and t
olerance ranges may be improved by merging this data-set with those of
other regions. The Spanish transfer function has strong affinities wi
th the African data-set and contributes important ecological data for
diatom taxa which are absent or poorly represented in the modern flora
of African lakes, and for which, in fossil material, there were previ
ously no good modern analogues.