Long-term trends in earthworm populations of cropped experimental watersheds in Ohio, USA

Earthworm communities in 7 experimental watersheds at Coshocton, Ohio were monitored twice annually (Spring and Autumn) for 5 years (1990-94). Samplin g by formalin extraction was undertaken along a transect running up slope w ithin the small (0.5-0.7 ha) watersheds that were used for row crop product ion under a variety of tillage practices. Six earthworm species were found: Aporrectodea turgida, A. trapezoides, Lum bricus rubellus, and Octolasion tyrtaeum present in all watersheds, plus A. tuberculata and L. terrestris with restricted distributions. In individual watersheds, earthworm density was lowest at 3 m(-2) (Autumn 1991) and high est at 397 m(-2) (Spring 1994), with overall watershed averages fluctuating between 55 and 247 m(-2). Tillage practice and crop type did not have, a s ignificant effect on earthworm numbers (p > 0.05). Equally, position along the transect was not significant for all watersheds, however, in some years (e. g. 1991) more worms were found towards the slope bottom. This suggeste d a link with rainfall but no significant correlations were found between r ecorded rainfall and earthworm numbers over the 5 year period. Additional s ampling during Spring 1994 using a combination of formalin extraction and h and-sorting revealed that some populations may have been underestimated by 75 % (this difference mainly due to the smaller species), but overall no si gnificant differences were recorded (p > 0.05) for biomass or density. The presence of deep burrowing species has a marked effect on rainfall infi ltration. Therefore, attempts were made to establish L. terrestris in one w atershed where it was previously absent. These proved unsuccessful, even th ough a number of techniques were used. This indicated that the distribution of earthworms within this system was not simply a factor of colonisation a bility.