The influence of environmental variables on the density of larval lampreys in different seasons

The objective of the study was to identify a subset of a set of twenty environmental variables which could explain variations in the density of larval lampreys (Geotria australis) in a south-western Australian stream. Generalised linear modelling, assuming Poisson distributions for the larval counts, led to a different model for each of the four seasons, with variations in larval density being explained in each season by a combination of between five and eight environmental variables. The influence of stream region also had to be taken into account in the model for winter.Four environmental variables (substrate organic material and chlorophyll a, macrophyte roots and low-angle shading) were present in three of the four seasonal models. A further six variables (water depth, substrate depth and profile, medium-sized sands, light intensity, and the presence of an eddy) were each found useful for two models. Two variables (current velocity and substrate profile) were each retained in one model. Eight of the twenty variables were not required for any of the seasonal models. The importance of organic material, shade, eddies, current velocity substrate particle size and a sufficient depth of substrate in our models agree with the largely subjective assessments of larval lamprey habitats made in the field by many previous workers for other lamprey species in diverse geographical localities.Our finding that larval density increased with increases in organic material and unicellular algae in the substrate and with shade, contrasts with the results of a different model based on data collected in a northern European stream. These differences can be related to our use of a more rigorous and comprehensive sampling regime and a more appropriate form of statistical analysis.