J. L. HARRISON 
Table 2 
176 
Infestation of marked rats captured after one and two days freedom 
For each species, or group of species, are shown the numbers infested and not infested 
with chiggers of any kind, and the proportion not infested. For comparison the “ expected ” 
numbers and proportion are shown in italics. The “ expected ” proportion is, for the one-day 
rats the square root of the observed proportion of the two-day rats, and for the two-day rats 
the square of the observed proportion of one-day rats. The numbers are compared by a 
chi-square test the result of which is shown together with the probability of getting such a 
value of chi-squared by chance. 
One day exposure. Two day exposure. 
Species of rat 
Inf. 
Not inf. 
Proportion. 
Inf. 
Not inf. Proportion 
chi 2 
P 
not inf. 
not inf. 
R. exulans 
11 
62 
.85 ... 
20 
27 
•57 
17.5 
55.5 
.76 ... 
13 
34 
72 . 
. 8.4 ... 
0.14 
R. r.jalorensis 
20 
37 
.65 ... 
21 
25 
•54 
15 
42 
.73 ... 
26.5 
19.5 
.42 . 
. 4-9 ••• 
0.09 
R. zvhiteheadi ... 
3 
16 
.84 ... 
3 
6 
.67 
3.5 
15.5 
.82 ... 
2.6 
6.4 
.71 . 
. 0.17 ... 
0.9 
Others ( R.r . argentiv- 
enter, R. rajah , mulleri 
and sabanus ) 
13 
38 
•74 ••• 
12 
16 
•57 
12 
39 
.76 ... 
12.5 
15.5 
.55 .. 
. 0.15 ... 
0.9 
The central value of the expected infestation during two nights may, therefore, fairly be 
taken as midway between the values for the separate nights, i.e., at about 10 a.m. on the day 
before being brought in. The best estimate of the mean feeding time is thus the mean of the 
observed times plus half a day. The mode is estimated by dividing the modal group in 
proportion to the adjoining groups and adding half a day. The 95 per cent, range is taken 
as above. 
Wild rats 
Unmarked rats of unknown exposure presumably bear a mixture of the remains of 
successive infestations spread over as many days as the maximum feeding time of the chiggers 
concerned. Two methods of estimating the true feeding time have been used. 
1. The average mode. — Fig. 3 shows a histogram of the apparent feeding times of chiggers 
of the species E. ( Walchiella ) oudemansi from a wild caught specimen of Rattus bower si. 
Evidently this is not an example of random distribution about a mean, but of distribution about 
two modes at days 4 and 11, or possibly three at 4, 9, and 11. Presumably the population of 
mites on the rat represents two or three separate attacks and the peaks in the distribution 
represents the modes of times of the attacks. 
If a sufficient number of examples were taken presumably all possible apparent modal 
feeding times would be represented, from zero to the actual modal time, in equal abundance. 
The mean of all such modal values, therefore, should be equal to half the true mode. 
In practice a “ mode ” has been scored for each peak which is at least three mites greater 
than the values before and after. Day one has been counted as a mode when the score was 
three greater than on day two. The mean of all these modal values has been calculated and 
doubled to give an estimate of the true mode. 
A rough estimate of the maximum has been made by adding the mean of the observed 
maxima rat by rat, to the mean of the modal values (i.e. half the estimated mode). The 
reason for this is that only rats bearing a reasonable number of mites were kept over water. 
STUD. INST. MED. RES. 
