A combination of the figures from the three samples listed above provides an intermediate "smoothed" figure that can be accepted with somewhat more confidence than any one of the separate censuses because it is based on more records. The combined ratios indicate a total of 105 racers in the headquarters field area. The figures obtained in the different sampling periods, and the census figures derived from their ratios are shown in tables 22 and 23. Differences from year-to-year in the census figures for any one area show no consistent trends and their variation is similar to that shown in different samples for the same season. Probably populations were fairly stable throughout the periods involved. If such stability is assumed, the samples from different years may be combined, and the composite figures derived from them may be accepted with more confidence. For the headquarters field area, for instance, 419 records of racers were gathered in all the preliminary sampling periods of the seven seasons involved; the records totalled 263 for all the secondary sampling periods, with 70 recaptures in secondary periods of the snakes recorded in the corresponding preliminary periods. A population of 75 racers is indicated--1.9 per acre. Corresponding figures for the northeast field area are: preliminary samples 453, secondary samples 163, recaptures 39, calculated number 135 (2.82 per acre). For the Rockefeller Tract the figures are as follows: Preliminary samples 807, secondary samples 476, recaptures 126, the ratio indicating a population of 153, or 1.11 per acre. These figures represent the number of adults present in early summer when the population is near its annual low point. The first-year young, excluded from this census because they cannot be caught in representative numbers, perhaps approximate the number of adults, in May, so the figures obtained would need to be approximately doubled to be representative of the entire population. By late summer the adults, and especially the yearlings, have undergone substantial reduction in numbers, but in late August and early September the hatching of a new crop of young increases the population to its annual maximum. The maximum numbers probably are about three times those obtained by censusing adults in early summer. The peak population of late summer or early autumn is estimated to consist of hatchlings, comprising somewhere near 50 per cent; adults, comprising a little more than 25 per cent; and yearlings comprising a little less than 25 per cent. Table 22. Captures Recorded and Population Calculated From Them on Hilltop Grassland Areas of Rockefeller Experimental Tract and Adjacent Reservation in Four Different Years