Sun angle, site selection, parameter choice, and generalizing UV effects in freshwater benthic ecosystems (Reply to the comment by Donahue and Clare)

Donahue and Clare (1999) provide a number of criticisms of our study of the ecological effects of ultraviolet (UV) light in a Tennessee stream (Hill et al. 1997), asserting that errors in experimental design, site selection, and parameter choice prevented us from detecting potential UV effects on periphyton and grazers. Here, we examine the merit of these criticisms and reemphasize the potential variability of UV effects in freshwater benthic communities. In their first criticism, Donahue and Clare question the efficacy of the Plexiglas shields at low sun angles. They calculate the horizontal displacement of the UV shadow on the streambed from the Plexiglas (resulting from seasonal or daily variation in solar azimuth and elevation), arguing that this displacement would have caused significant exposure of substrates directly below the Plexiglas and significant shadow effects on adjacent control areas. We agree that the UV shadow would have been displaced at low sun angles. However, this displacement was inconsequential for two reasons. First, we countered the northerly displacement of the UV shadow that occurred during the autumn experiments when the sun was low in the south by simply placing the experimental substrates in a northerly direction on the streambed, using the visible shadow of the Plexiglas-supporting steel rods as a guide to the northerly displacement of the UV shadow. Second, the daily west-to-east movement of the UV shadow would have had little effect because the experimental substrates were much narrower than the UV shadows. All periphyton and snail response variables were measured on these substrates, which consisted of groups of tiles 10 cm wide (16 tiles, each 2.5 3 2.5 cm, arranged 4 3 4) centered at the east-west midpoint of the Plexiglas shields. Using relevant angles of solar altitude and azimuth and a refractive index of 1.358 for UV (specifically, that of 303 nm), we calculate that on 11 November, when the potential effect of hourly changes in sun angle is greatest in this study, the 10cm-wide groups of tiles 15 cm below 403 40-cm Plexiglas shields would have been completely shaded from UV during the 6-h period surrounding solar noon. This is the period when direct sunlight fell on the stream in autumn; direct sunlight was blocked by adjacent trees and the streambank earlier and later in the day. Moreover, it is the period during which .90% of the potential daily UV-B irradiance occurs on any day of the year. Tiles were 15 cm below the Plexiglas in three of the four blocks in our experiments. The group of tiles 35 cm below the shield in the remaining block would have been shaded 75% of this time. The percent UV-B screened out by the Plexiglas shields can be calculated with time-specific UV-B irradiances obtained from UV-B: Version 3.00 (Fiscus and Booker 1994): 78% of the potential UV-B radiation impinging upon sub-