TOf. METHODS, ,1\D TECHMQIFS 



which flushes would occur in relation to the various cover tvpes. if grouse exercised no choice 

 of them, was a necessary preliminary step. 



For the purpose of the ln\cstif.'ati(iii. the ''no choice" pattern, witli one exception, was set 

 up on the hasis of the amount of ear li t\ pe ])resent. Three of the studv areas were located 

 in a haliitat t\pe characterized 1)\ ahaiuloned farm.*, so the woodland, though sizeable, was 

 surrounded by open fields. Since these are not used by grouse except along the edges border- 

 ing the woods, an arl)itrar\ limit of TOO feet from existing wood or brushland was taken as the 

 extent of open land that should l)e included in figuring the amount of each type present. It 

 was also ob\'ious that the habit of frequenting the edges, rather than the centers of large 

 blocks of cover, excluded other areas from the birds' use. Therefore, a measure of the amount 

 of each type available was better than the total amount for determining the pattern of 

 occurrence which would be associated with no choice regarding cover type. No practical 

 method existed, however, for evaluating availability. 



The reliability of the information was checked by two methods. If 20 per cent of an area 

 was composed of a certain type of cover and 25 per cent of the flushes during a ])eriod 

 occurred in that type, the validity of the difference was determined by the statistical method 

 of chi-square*. More detailed problems, such as those dealing with difference in cover type 

 usage between areas, seasons or years, were checked by analysis of variance. 



Determining Food Relationships 



The items of food upon which the grouse subsists were determined by an examination of 

 the gizzard and crop. The birds collected during the course of the Investigation were taken 

 by various means and by a variety of individuals. Satisfactory sampling of the grouse popu- 

 lation has de|)ended on the cooperation of sportsmen, naturalists, game |)rotectors. special 

 collectors and the public in general. An attemjU was made to maintain a uniform flow of 

 birds to the Research Center in order that continuous infortnation by regions during all 

 seasons throughout the period of studv. with a good re|)resentation of both sexes and different 

 age groups would be |)rovided. 



Through the cooperation of the above-mentioned groups, birds have been received which 

 were found dead in the field, hit bv automobiles and trains, those which have flown into 

 houses or wires, and others killed by predators a.s well as a considerable number taken 

 during the regular open season. The backbone of the continuous study, however, was formed 

 by collections made by shooting specified numbers of birds from designated areas at pre- 

 scribed times. For this purpose, special collectors were hired as needed. Certain game 

 protectors were also assigned to the job. The standard methods of procedure as developed 

 and described"*' "^ by the Section of Food Habits of the U. S. Fish and Wildlife Service, were 

 followed in making the analvses. Food items difficult to identify were referred to specialists. 



Simple but important refinements in the method of determining the average volume of each 

 food item, as obtained from a number of analyses, were adopted. Previously, the volumetric 

 percentage of a single item in the total diet has been figured by first finding its percentage 

 in each individual bird, then averaging these to secure the jjercentages for all birds. This 

 procedure is, of course, mathematicallv unsound. For example, one bird may eat only two 

 beechnuts, rejjresenting 100 per cent, while in another, the same tiumber makes up oidy 10 per 

 cent of the volume. Averaging these, four beechnuts become rin |)er cent of the total volume. 

 Obviously this is not true. 



• For tcrminolngy anil method* of anntvliiit. *rr «li«rii»^ttiii of ITur o( Itioiiirliiral Aiuil*.;". [•. 71H. 



