There are not many studies that have evaluated this issue, so this is a valuable proposal. The background for the proposal was adequate. This study has the potential to evaluate carcass analogs, provided a suitable experimental design can be implemented. The development of carcass analogs represents a new technology that deserves investigation in controlled field studies before the region commits to wholesale acceptance.

In general, the proposal does a good job of relating the study to the general issue of deliberate nutrient enrichment to boost stream productivity, although the early work of C. E. Warren and colleagues at Oregon State University on nutrient enrichment of streams is often overlooked and should be reviewed by project sponsors. The practice of releasing salmon carcasses from hatcheries is widespread, but there are considerable logistical problems with deploying large numbers of carcasses throughout a stream network. The recent development of carcass "analogs" has been suggested as a much more tractable method, with the additional advantage of being able to deploy the material at the desired time and place -- not just when fish are available from a hatchery. Relatively few studies have monitored the biological effects of deliberate carcass releases, and with this new technology the effects remain largely unknown. The proposal does not explain what carcass analogs are (pelletized, pasteurized fishmeal derived from spawned-out hatchery salmon), and of the five assumptions about their advantages given, only assumption 3 (easy to transport) and assumption 5 (stable supply) should be taken at face value. The others (pathogen-free, closely mimic nutrients from natural carcasses, and similar breakdown rate) should be tested.

The proposal describes how carcass analogs have been deployed in the Wind River watershed in 2005, but does not mention any results. The claim is made that the Pearsons et al. (2003) study of carcass analog enrichment of a Yakima River tributary "restored food pathways by direct consumption and food chain enhancement"; however, in a recent presentation these authors have further stated "Except for an initial increase in growth approximately 6 weeks after analogs were stocked, we detected no effect of analogs on either growth or abundance of trout." Two other important references are omitted: Sanderson and Kiffney's (2003) progress report on carcass analog additions to streams in the Salmon River basin, and S. Claeson's M.S. thesis at OSU on experimental whole carcass effects on food webs in the upper Wind River.

The objectives are worthwhile and the proposal does a good job of covering the bases with regard to biological response -- water chemistry, periphyton, benthic invertebrates (although it is odd that only grazers will be analyzed for stable isotopes), resident fishes, and contaminants.

Using a predetermined range of carcass analog densities is a good idea, since the Yakima study of Pearsons et al. did not appear to have detected sustained trophic enrichment. One of the most important questions the proposal does not address is how the amount of natural spawning by Chinook and steelhead will be factored into the analyses. Although the proposal does not contain a map of the study streams, they appear to be located in the vicinity of the Carson hatchery. Thus, it seems possible that there will be some natural spawning in the study streams (the proposal does not specify if sites will be located above barriers to anadromous species). If natural spawning is distributed unequally among the study sites it could confound the objectives of the research. If there is no salmon spawning at any of the study sites, the objective of the work is slightly compromised because the study will have taken place in streams where aquatic communities have not adapted to historical salmon spawning over time. The proposal does not justify why a 500m upstream control and 500m downstream treatment approach was selected, as opposed to treating an entire stream with carcass analogs and pairing sites with untreated control streams to the extent possible.

The methods for sampling the periphyton, aquatic invertebrates, and fishes are standard techniques and should work well. Surprisingly, fish species were not specified. How does fish community composition vary among streams?