the Dreyer Ranch route, bringing the number of species recorded on all runs of 

 all routes to 108, or 78% of the total. 



The Dreyer Ranch route sampled more water habitats, and also larger 

 reservoirs, than the four standard routes combined (table 5); thus, it is not 

 surprising that the five species recorded only on this route are all aquatic 

 or semi-aquatic. Of the 36 species "missed" by the four regular roadside routes, 

 25 (59%) are aquatic or semi-aquatic and 20 (56%) are migrants or summer visitors. 

 Many of these were seen only on or near Fort Peck Reservoir, the Missouri River, 

 or other large bodies of water not sampled by the four standard routes. Had 

 moisture conditions been more favorable in 1977, it is likely that more aquatic 

 species, especially migrants, would have been recorded, as most water areas 

 along these routes were dry at the time of peak spring migrations. Excluding 

 aquatic or semi-aquatic species, the four standard roadside counts sampled 91 

 of 114 remaining species, or 80%. It thus appears that the technique is quite 

 effective in sampling those species which occupy habitats available along the 

 routes. 



Seasonal Changes in Sample Abundances . As discussed earlier, the sample 

 abundances (a) obtained for each species by no means represent actual abundances, 

 but are related to the actual abundances (x) by some unknown function f(x)=a 

 which incorporates various species-, area-, and season-specific variables such 

 as conspicuousness, local habitat, timing in relation to the reproductive cycle, 

 group size, and animal size. While it would certainly be interesting to define 

 tiie functions f(x) for each species and for each month, this is not possible 

 without exhaustive effort far beyond the scope of this study. Further, defin- 

 ition of f(x) is not necessary to draw ecological conclusions about individual 

 species based on a, as long as one realizes that a is merely a species-specific 

 abundance index, and comparisons between species or between monthly replicates 

 are made with due caution. 



It may be argued that the samples obtained by this technique are quite dif- 

 ferent in terms of relative abundances and diversity than the actual community 

 being sampled, and hence synecological interpretations are unwarranted. However, 

 the same may be said of virtually every cofimunity sampling technique, from small 

 mammal trapping (where some species are more trappable than others) to plankton 

 sampling (where smaller-bodied members of the community are not sampled at all); 

 while the sample abundances differ from actual relative abundances in all cases, 

 many valid statements about community ecology have been made based upon such 

 data. We will thus treat the samples as communities in the following discussion, 

 while being fully aware of sample bias. 



These data allow comparisons of relative abundances between routes, if, 

 for example, the values ofa for ring-necked pheasant are consistently greater 

 on route A than route B, we can conclude that abundances within habitats sam- 

 pled are greater for route A than for route B, and that route A provides more 

 suitable habitat overall than route B. These kind of conclusions have been 

 drawn in the species Narratives. 



Plots showing monthly changes in a for each of the five routes for the 

 four standard routes combined have been shown for selected species. These plots 

 may be grouped into four broad categories, with respect to their ability to 

 clearly separate May-June runs of the five routes. (1) Some plots show no 



153 



