BR. J. ENT. NAT, I MSI ., 7: 1994 49 



that a particular deviation, in an observed data set, from the expectation based on 

 a scientific hypothesis, for example that the frequencies of typical and melanic 

 P. rhomboidaria should be the same in yew woodland and chalk grassland, could 

 occur by chance sampling error.) The comparisons tested, the chi-squarcd values, 

 degrees of freedom and corresponding probabilities are given in Table 2. 



The data from P. rhomboidaria and A. repandala confirm the findings of I 

 et al. (1993). In P. rhomboidaria, significantly more melanics were taken under 

 the yew canopy than in the chalk grassland. (In this context, significant is used as 

 a technical statistical term, meaning that the likelihood that a particular deviation from 

 expectation is less than one in twenty, if the usual significant level (P) of 0.05 is set.) 

 In A. repandata, the relative frequencies of the half-melanic, melanic and banded 

 forms were homogeneous in the two habitats, with the frequencies of all three being 

 higher under the yew canopy than in chalk grassland. However, the frequency of 

 the typical form when compared to the three darker morphs was significantly higher 

 in the chalk grassland than under the yew canopy. 



In the case of /. aversata, the frequency of the banded form was higher in the 

 yew traps than in the chalk grassland traps, but not significantly so. However, 

 the data are not significantly different from those obtained by Jones et al. (1993), 

 when trapping at Juniper Bottom, in 1990 (data given in Table 3 for convenience) 

 either under the yew (x^ =0.070; P>0.05), or in the chalk grassland (\, =0.700; 

 P>0.05). 



The frequencies of darker forms (half-melanic and melanic) of Ectropis bistortata 

 (Goeze) were significantly higher inside the yew woodland than in the chalk 

 grassland. 



Conversely, the frequencies of the darker grades (classes 3, 4 and 5) of A. 

 monoglypha, did not differ significantly from those of the paler grades in the 

 two habitats. In H. furcata, the frequency of full melanics in the yew traps was 



Table 2. Analysis of data, giving chi-squared (x 2 ) values, degrees of freedom (d.f.) and 

 probabilities (P) (ns = non-significant). 



r d.f. p 



P. rhomboidaria 



Typical vs melanic, yew vs grass: 14.40 1 < 0.001 



A. repandata 



All forms, yew vs grass: 



Melanic vs half-melanic, yew vs grass: 



Melanic + half-melanic vs banded, yew vs grass: 



Melanic + half-melanic + banded vs typical, yew vs grass: 



E. bistortata 



Melanic + half-melanic vs typical, yew vs grass: 4.500 1 <0.05 



/. aversata 



Typical vs banded, yew vs grass: 1.237 1 ns 



A. monoglypha 



Classes 1 +2 vs classes 3 + 4 + 5, yew vs grass: 0.002 1 ns 



H. furcata 



Typical + half-melanic vs melanic, yew vs grass: 2.157 1 ns 



9.352 



3 



<0.05 



1.610 



1 



ns 



0.064 



1 



ns 



7.869 



1 



<0.05 



