Oecologia (Berlin) (1986) 68:512-515 Oecologia 9 Springer-Verlag1986 Rainfall and decomposition in the chihuahuan desert W.G. Whitford, Y. Steinberger*, W. MacKay, L.W. Parker**, D. Freckman***, J.A. Wallwork****, and D. Weems Biology Department, Box 3AF, New Mexico State University, Las Cruces, NM 8803, USA Summary. We tested the hypotheses that rates of decompo- sition in a desert should be higher following single large rain events of 25 mm than evenly spaced 6 mm events and that supplemental rainfall should result in higher popula- tions of soil biota. There were no significant differences in mass losses of creosotebush, Larrea tridentata, leaf litter on plots receiving water supplementation and no added water. On some sampling dates, there were higher mass losses in the 6 mm. week- 1 treatment. Weekly rainfall pro- duced higher coefficients of variation in mass losses than the other rainfall regimes. A single event pulse compared with weekly pulses of rainfall during the normal "dry" period resulted in no differences in mass losses. Microarth- ropods and nematodes exhibited numerical responses to supplemental rainfall but the litter microflora did not. These studies provide direct experimental evidence that the conventional wisdom linking decomposition to rainfall in deserts is wrong. The studies also suggest that the effects of litter fauna on surface litter decomposition are minimal; therefore, future studies should focus on activites of the microflora. Conventional wisdom holds that deserts are water limited systems and that their biological processes are triggered and maintained by rainfall (Noy Meir 1973; Louw and Seely 1982). Statements to the effects that decomposition in deserts must be limited to short periods when soil and litter are moist (Noy-Meir 1973) are generally accepted hy- potheses despite some evidence that decomposition does not vary with rainfall as predicted by general climate models (Whitford et al. 1981 ; Elkins et al. 1982). In studies of de- composition in several North American deserts, we found Offprint requests to: W.G. Whitford Present addresses: * Y. Steinberger, Department of Biology, Bar-Ilan University, Ramat Gan, Israel ** L.W. Parker, Westbridge Research Group 3770 Tansy St., San Diego, CA 92121, USA *** D. Freckman, Department of Nematology, University of Ca- lifornia, Riverside, CA 92502, USA **** J.A. Wallwork, Department of Biology, Queen Mary Col- lege, University of London, Mile End Road, London E1 4NS, England a better correlation between decomposition and long-term average rainfall and the season when rainfall is most likely to occur, than with the actual rainfall occurring during the study period (Santos et al. 1984). These findings led us to reexamine the trigger-pulse-reserve paradigm proposed by Noy Meir (/973) in terms of its' applicability to decomposi- tion processes in desert ecosystems. Investigators working in North American deserts have argued that high quantity rainfall events, e.g. events of greater than 25 mm are needed to trigger biological activity such as ephermeral plant germination, and that low quan- tity events such as 6 mm are relatively ineffective as triggers of biological pulses (Beatley 1974; Tevis 1958; Went 1949). We hypothesized (1) that decomposition processes should respond like ephemeral plant populations, i.e. that decom- position should be higher following a single 25 mm rain event than after four evenly spaced 6 mm events, and (2) that both rainfall patterns should result in higher decompo- sition rates than occur after natural rainfall. Methods We established nine plots on an area where creosotebush, Larrea tridentata occurs as a monotypic dominant shrub. The study site in the northern Chihuahuan desert 40 km NNE of Las Cruces, NM receives a long-term average rain- fall of 250 mm-yr -1. The "rainfall supplemented" plots were irrigated with sprinklers that delivered water above the shrub canopy to mimic natural rainfall as closely as possible. The irrigated plots thus received the natural rain- fall plus the scheduled supplement. Three plots received 25 mm every four weeks, 3 plots received 6 mm per week and 3 plots received no additional water. We measured mass loss of organic matter and estimated population sizes of biota in open bottom aluminium screen cylinders that initially contained 20 gins of creosotebush (L. tridentata) leaves. We also measured mass loss in fiber- glass mesh bags. The creosotebush (Larrea tridentata) leaves were collected from shrubs in April and May and air dried in outdoor drying racks. Subsamples of leaves were oven dried to a constant weight to obtain a correction for the air dried leaves. Larrea tridentata leaves have a C: N ratio of 26.7 and an initial lignin content of 10.63% (Schaefer et al. 1985). The experiment was started in June 1981 and a second set of cylinders was placed in the field in December 1981 and collected at 3 month intervals for