Hi John Unless I'm missing something, I'm not sure there's much utility in comparing initial weight to final weight. It would make much more sense to compare initial weight between the two treatments, and final weight between the two treatments. Comparing initial weight may give us an idea whether the treatment environments influence the condition of the fish at hatching.
Another useful measurement would be that of growth rate, which effectively combines your before and after mass information into a single value. Since fish don't exhibit linear growth, you would use a formula for specific growth rate: G = 100 x [log(Wf) - log(Wi)] x (t^-1), where Wf = final weight, Wi = initial weight and t = time between the two periods (look in e.g. Hojesjo et al. 1999, Journal of Fish Biology, 55, 1009-1019, for more info). One potential problem would be your method of analysis, of course. Since you have 40 fish per treatment in each stream those would probably be considered pseudoreplicates. I think the easiest thing to do would be to average data for each stream, and then your treatment groups would be composed of n=8 replicates (one value per stream). There are no doubt other alternatives, though, which a statistician might be able to tell you about. It looks as though Bill's papers would provide good examples of how to work with these data, anyhow. Hope that helps Jack On 31 May 2012 18:48, Resetarits, William <william.resetar...@ttu.edu>wrote: > John, > > Their are lots of examples out there, but here are two papers describing > experiments that are very similar to yours in many elements of design and > execution, both involving fish, experimental streams, and comparisons that > involve initial and final measurements on groups of individuals that are > not individually identified. > > Resetarits, W. J., Jr. 1995. Limiting similarity and the intensity of > competitive effects on the > mottled sculpin, Cottus bairdi, in experimental stream communities. > Oecologia 104:31-38. > > Resetarits, W. J., Jr. 1997. Interspecific competition and qualitative > competitive asymmetry > between two benthic stream fish. Oikos 78:429-439. > > Good luck, > Bill > > On 5/31/12 10:57 AM, "John Winkowski" <john.winkow...@gmail.com> wrote: > > Hello fellow ecolog'ers, > > I am an MSc student studying environmental effects on phenotypic > development and fitness of > juvenile Atlantic salmon. > > I conducted a study where I incubated salmon eggs in 2 different > environments, measured a few > phenotypic traits, and placed equal numbers of fish from both incubation > environments in "semi- > natural" stream channels to measure survival and growth in the first 45 > days post-emergence. Thus, > fish from both treatment groups were competing in the stream channels with > only natural prey items > available. > > The fish were too small to individually mark them at the onset of the > experiment, so I could only > mark them to treatment with VIE tags. I was wondering if anyone had any > suggestions on how to > analyze such group data. Is it legitimate to compare average initial > weight with average final weight? > > I have the lengths and weights at the onset of the experiment and at the > end of the experiment. 40 > fish from each treatment group went into each stream (thus, 80 > fish/stream, with 8 replicate > streams) and the streams were closed systems. At the end of the experiment > I drained each stream > and measured (length and weight) the surviving fish. > > Any insight would be excellent! Thanks so much. > > Cheers, > > John Winkowski >
