recent years (Kicker 1948). Tlie method deiieiuls 

 first on capturing a fair sample of individuals in a 

 unit area, marking them in some distinctive manner 

 (Ecol. 37, 1956: 665-689), releasing them for redif- 

 fusion over the area, then after a short interval, re- 

 trajjping the area. The ratio of marked individuals 

 reca])tured to the total numher marked sliould theo- 

 retically be the same as the total marked and un- 

 marked animals captured during the second trapping 

 is to the total population or : 



Total ]ioini]atioii 



il numher marked 



marked individuals 

 X total captured 



Other formulas make use of accumulating totals of 

 marked and unmarked individuals during successive 

 periods of trapping (DeLury 1958). 



The greater the percentage of the population 

 marked and subsequently reca])tured, the greater is 

 the accuracy of the calculations. However, there are 

 several possible, uncontrollable sources of error : un- 

 equal mortality of marked compared with unmarked 

 individuals ; dispersal of individuals out of the area, 

 influx of animals from outside : increase by reason of 

 reproduction, marked animals not becoming randomly 

 distributed among the unmarked ; marked animals 

 being recaptured with greater or less ease than un- 

 marked ones : marks being lost or not reported, and 

 so forth. Some of these possible errors can be cor- 

 rected statistically, and a considerable body of litera- 

 ture has accumulated describing means of so doing 

 (see especially Bioiiietrika since 1951). 



CAPTURE PER LLMT-EFFORT 



In a closed or stabilized population, when 

 the same time, traps, and effort are employed to cap- 

 ture or count individuals in the same area at difYerent 

 times and there is no loss or increment in the 

 original population, and weather and other conditions 

 remain the same, the number of new individuals cap- 

 tured or discovered witii each subsequent effort be- 

 comes less and less, and should eventually reach zero. 

 When the number of new individuals captured per 

 unit of effort is plotted against the cumulative num- 

 ber of animals captured, a straight line results. A 

 line thus derived from a few catches may be extended 

 to zero, and the total population of animals in the 

 area determined (DeLury 1947, Zippin 1958). A 

 variation of this method is to use the increasing per- 

 centage of marked animals in the total number cap- 

 tured at successive intervals of time, as the increase 

 in these percentages follows a definite trend that 

 would eventually include the total population (Hayne 

 1949). 



ai'PI.icahon to am.mal groups 



Tiie more conspicuous diurnal mammals are 

 commonly censused by cruising or drives, but noc- 

 turnal forms, es])ecially mice and shrews, usually have 

 to be trapped (but see Kmlen ct al. 1957). Snap or 

 kill traps are commonly used. When set in a variety 

 of microhabitats they quickly gather specimens to 

 show the species composition of the community. An 

 early attempt at estimating abundance was expressed 

 in terms of the number of individuals caught per trap 

 per night. At the same site, more animals are usu- 

 ally caught during the first night than during later 

 nights: 10 traps set for 10 nights will not capture as 

 many small mammals as 100 traps set for only one 

 night, although 100 trap-nights are involved in both 

 instances. When the trapping procedure is standard- 

 ized as to location in community, number of traps 

 used, interval between traps, length of trap lines, 

 number of nights trapping, and so on as has been 

 done in the Xorth American Census of Small Mam- 

 mals (Calhoun 1956), it is possible to follow changes 

 in relative abundance from year to year. It is not 

 ])ossible to relate such data to the absolute number 

 per unit area unless the home range of each species 

 in each locality is known (Stickel 1948). 



The next advance in censusing technique was to 

 confine the location of kill traps to a small area, usu- 

 ally an acre (0.4 hectare). Enough traps, a hundred 

 or more, are included to saturate the area to the end 



100 



200 400 600 800 1000 1200 

 TOTAL CATCH TO DATE, m 



FIG. 42 Total population {K = 1170) calculated by extension 

 of a straight line through data on successive catches per unit 

 effort, C(/), plotted against the accumulating total catch, K(i) 

 (from DeLury 1947). 



Measurement of populations 35 



