Biomass measurements are important for answering questions about trophic dynamics in aquatic systems. Measuring dry weight for biomass is a destructive process, but specimens may be needed for other processes (e.g. stable isotope analysis) or for long-term preservation. Length-mass regression is a non-destructive process that is widely used for calculation of biomass. However, length-mass regressions have not been calculated for many macroinvertebrate taxa common to the Middle Rio Grande, an aridland river in New Mexico. Potentially, individuals of the same taxa may be the same length but have different dry mass in the Rio Grande compared to their mesic-system counterparts due to environmental differences (e.g. flow variability, geology). This may hold true for other aridland river systems, posing the need for new equations to be produced. Macroinvertebrates were collected from 5 sites along the Middle Rio Grande from summer 2007 to 2008.  Abundant taxa were measured and sorted into classes by body length and head width, then oven-dried and weighed. Regression analyses were performed on most abundant taxa calculating allometric equations for: Corisella (Corixidae), Physa (Physidae), and Erpetogomphus (Gomphidae). Also, allometric equations for Tricorythodes (Tricorythidae) were compared among our collections from the Middle Rio Grande, and other equations from previous studies conducted in Virginia and Idaho.

Results/Conclusions

Using the equation M=aL^b, head width and body length both provide reliable estimates of biomass for both adult and nymph macroinvertebrates sampled. Amongst all of the taxa tested, regressions do predict biomass given a length measurement shown by the strong correlation factors (r²>0.75). For example, the regression between head width and dry mass for adult Corisella was strong (r²=0.79), as was the regression for Physa shell length (r²=0.92) and shell width (r²=0.87). The regression equations developed here will be able to be used to estimate biomass of these taxa (p<0.001). Furthermore, allometric equations developed from samples from the Middle Rio Grande were similar to those calculated elsewhere (p<0.05).