l)iii-])()iiUe(l. Roadsidi' counts, calling-male transects, 

 indices derived from i>opiilation structure, kill rec- 

 ords, and a variety of other procedures are used to 

 inventory upland game species (Hickey 1955). 



I'or determining populations of smaller sjiecies 

 during the nesting season, the spot-map method is 

 commonly used and censuses thus obtained are prob- 

 ably reliable within plus or minus 10 per cent, if they 

 are carefully made (Kendeigh 1944). A sample jjlot 

 of uniform vegetation of at least 10 hectares {2S 

 acres) is marked out in a grid with numbered stakes 

 or tree tags at intervals of not over 50 meters, or the 

 stakes may be placed along a trail. At least five, 

 preferably more, daily counts of singing males, fe- 

 males, and nests are made at suitable intervals 

 throughout the nesting season. Kach time a bird is 

 observed it is marked on a map of the plot. At the 

 end of the season all the spots at which a species was 

 observed are ])laced together on one map. Since in- 

 dividual birds are observed most frequently in the 

 vicinity of their nests and within their territories, the 

 spots fall naturally into groups so that each group 

 indicates the presence of a breeding pair or at least 

 a territorial male. Counting the number of groups of 

 spots for each species gives the total population for 

 the area. For the large predators, gallinaceous birds, 

 or wide ranging species, census plots of much larger 

 size are necessary than for the smaller song birds, so 

 that procedures must be adjusted to the conditions 

 of the habitat and the species involved. For detailed 

 studies of small populations, the birds should be 

 handed and color-marked for individual recognition 

 (Hickev 1943). 



Foliage arthropods 



In order to determine the insect and spider 

 composition in the herb, shrub, and tree strata of a 

 forest, use of a variety of collecting methods is de- 

 sirable : net sweepings. light traps, bait traps, ad- 

 iiesive snares, and the like (HofTmann ct al. 1949, 

 Morris 1960). Some of these methods may be made 

 semiquantitative to show relative abundance, but 

 there is considerable difficulty in converting the data 

 obtained into absolute abundance. 



The use of the sweep net can be standardized to 

 give useful and comparable estimates of population 

 densities (Carpenter 1936). A series of 48 strokes 

 of the net through the upper level of the herb stratum 

 synchronized with one's pace so that successive 

 strokes do not hit the same plants gives approxi- 

 mately the same number of individuals as one would 

 find on the herbs covering one square meter if all 

 could be captured. The net should have a diameter 

 of 33 cm (13 in.), the strokes should be about one 

 meter long (.Shelford 1951a), and comparative sam- 



ples should be taken at appro.ximately the same time 

 of day (Adams 1941). A similar number of strokes 

 through the shrub foliage may be used, but the con- 

 version to number of individuals per S(|uare meter 

 depends on the extent and uniformity of the shrubs 

 that cover the ground. Inaccuracies involvetl in 

 sweep net sam|)ling are the result of variations in the 

 activity of the insects and s])iders ])roduced by 

 changes in teni])erature, wind, and humidity ; varia- 

 tions in position of the insects on the plants and hence 

 exposure to capture ; insects taking flight in advance 

 of the collector; variations in the height of the herbs; 

 and variations in the length and rapidity of the 

 strokes (DeLong 1932. Hughes 1955). Differences 

 between sexes and species in behavior and life his- 

 tory will also cause variations in the sampling effec- 

 tiveness. 



Tests on the reliability of population estimates of 

 single species based on the sweep-net method, made 

 by comparing the results of two different workers in 

 the same woods at the same time, showed an agree- 

 ment within 100 per cent in only 36 per cent of com- 

 parisons between single weekly collections, but in 74 

 per cent of comparisons between averages of weekly 

 collections taken over the entire summer (Graves 

 1953). This would indicate that variations in popu- 

 lation estimates obtained by sweep-net samples are 

 not significant unless a good series of data is ob- 

 tained, and only then when differences between aver- 

 ages amount to more than 100 to 200 per cent; i.e., 

 when the larger population is at least 2 or 3 times 

 the size of the smaller. .Xctually. variations in popu- 

 lation size of the same species of insect or spider at 

 different times or in different communities m.iy 

 amount to several hundred per cent, and hence the 

 sweep net method is useful for quantitative studies. 



Sampling of arthropods in the tree canopy is 

 more difficult. In the absence of wind, small trees 

 can be jarred or shaken so that released animals fall 

 on a cloth spread beneath. With proper e(|uipment, 

 trees may be fumigated with such poisonous sprays 

 as DDT so that the dead insects fall onto cloths 

 spread below. To put the data on a comparative 

 basis, the volume of the space occupied by the foliage 

 may be measured or estimated, and the number of 

 individuals per cubic meter calculated. In deciduous 

 forests of eastern Xorth .America the tree canopy is 

 commonly about 10 m thick. A useful standard for 

 comparison with the numbers per square meter of 

 ground, herbs, and shrubs is the number per 10 nr''. 



With taller trees, samples of the foliage for visual 

 counting of the immature stages of arthropods pres- 

 ent may be collected with the aid of aluminum pole 

 pruners and extension ladders, or from trestles or 

 platforms. Foliage samples, especially of coniferous 

 species, should consist of entire branches or longi- 

 tudinal halves, since arthropods may vary in abun- 



Measurement of populations 37 



