Analysis of Variance 
Since this is a statistical analysis of correlated temperature and hatching 
data, the class or set of things being discussed must be clearly understood. 
The term hatching period applies only to that part of the infestation that 
survives to become a problem of control, damage to crop, and perpetuation of the 
infestation. This has been called the end infestation and is made up of all 
those survivors that have escaped all factors that tend to destroy newly hatched 
nymphs. 
In table 5 is the analysis of variance for the testing of the significance 
of the regression for each of the four species included in this study. The F 
values for the temperature regression are highly significant, being well above 
the l-percent level (99 to 1 odds) in every case. 
It can be concluded that daily maximum temperatures are a measure of heat 
and have an additive effect on the rate of embroyonic development. Laboratory 
tests in hatching eggs at controlled temperatures have already proved that heat 
has this effect. This finding permitted the assumption that the effect of heat 
on hatching at any time is independent of the amount of heat occurring at any 
other time. The analysis of variance has now given high significance to the 
assumption that use can be made of orthogonal polynomials for predicting the 
hatching period of the end infestations of grasshoppers. 
According to Hee cetimn "This is done by the application of Fisher's device 
which gives a regression curve that shows the effect on hatching of a unit change 
in a given meteorological element at any time during the growing season." For 
this problem the number of days after April 30 when 70 percent had hatched has 
been substituted for yield and the unit change in a given meteorological element 
is 1 degree of average daily maximum temperature for a 5-day interval. 
These regression curves are shown in figure 2 for each of the four species. 
Houseman>/ gives a method for interpreting these curves. This involves the 
testing for significance of the coefficients, the Ta's, which measure various 
characteristics of the temperature distribution during the periods of time in- 
cluded in the curves. The regression coefficients are shown in table 6a and 
their significance is tested by using the t values in table 6b. 
The Taj is a measure of the effect of the total temperature for the periods 
indicated for the curves in figure 2. According to table 6b, the Tag regression 
coefficients are statistically significant, which means that, for the four 
species, an increase, in the total temperature shortens the time after April 30 
when development of the infestation takes place. This is to be expected. 
The second measure, Ta,, is proportional to the average increase or decrease 
of temperature per 5-day interval during the periods included in the curves. The 
Ta, regression coefficients are statistically significant for both sanguinipes, 
especially the early hatching ones, and bivittatus, but not for differentialis 
and femurrubrum (table 6b). 
4/ Houseman, E. E. Methods of computing a regression of yield on weather. 
Iowa Agr. Expt. Sta. Res. Bul. 302, pp. 863-904. 19h2. 
b/s Ebide 
