io8 DAVID JENKINS AND ADAM WATSON 



season that a number of long-dead corpses are found but it is encouraging 

 that the number of carcases overlooked earlier have been few in relation to 

 those found soon after the grouse died (Table VII). 



The cause of death can be ascribed to one of three classes in the field. These 

 are (i) predation; (ii) accident; (iii) 'others'. Grouse killed by predators usually 

 lose some feathers and are often partially eaten (but here post-mortem 

 changes and scavengers can cause confusion). In the case of accidents the 

 grouse is found with some injury below an overhead wire or by a fence, or 

 entangled in wire or caught in a snare or gin. Occasionally birds are struck 

 by motor vehicles or poached. The cause of death of birds which have not 

 died through violence is classified under 'others'. This includes disease. 



We do not know what proportion we find of the total number of grouse 

 that die on the study area (but see Jenkins (1961). Most of the heather is short, 

 nine-tenths being less than 30 cm (12 in.) high. The dogs point at freshly 

 dead grouse and we can often see these at a distance. Predation usually results 

 in some feathers lying on the heather and these detached feathers, ghnting 

 white, can be seen from 100 m or more away. 



Possibly we find a higher proportion of predation deaths than of birds 

 that died from other' causes but it is clear (from the relatively small number 

 of corpses we fmd long after death) that most of the deaths in the latter 

 category occur in the spring and summer. Such carcases are often concen- 

 trated along streams and in other green places. It is noteworthy that we find 

 dead birds in some months and not in others, and that we have found most 

 dead birds in each category at similar seasons in different years. Our findings 

 probably indicate the proportions dying in different seasons, if not the exact 

 numbers dying, and they suggest that relatively few grouse die from 'other' 

 causes in the autumn and winter, the seasons which which we are concerned 

 in this paper. 



RESULTS 

 The count figures (Table II) suggest that losses from the population occurred 

 in two or three distinct stages each autumn and winter. For convenience 

 these stages may be classified as (i) between the August count and the last 

 count in December (the autumn decrease), and (ii) between January and 

 March or April (late winter decrease). Comparable figures for percentage 

 change were as follows : 



Autumn Late winter 



1956-7 — 24 



1957-8 50 24 



1958-9 49 43 



1959-60 37 — 



