Colour-pattern variation in relation to habitat in the grasshopper Chorthippus brunneus (Thunberg)

Ecological Entomology (1 979) 4,249-257

Colour-pattern variation in relation to habitat in the grasshopper Chorthippus brunneus (Thunberg)

PETER D. GILL Department of Genetics, University of Nottingham

ABSTRACT'. 1. Chorthippus brunneus were collected from twenty-two different sites in Nottinghamshire, Derbyshire and Essex.

2. Discrete characters of the head, pronotum and forewings and hind femora were scored. The results were compared with the types of habitat from which the grasshoppers were collected.

3. Associations between striped and mottle-winged varieties were found. 4. The relationships between colour and habitat were examined.

Introduction

The grasshopper Chorthippus brunneus (Thunberg) is extremely variable in colour and wing pattern; most are a shade of brown or grey. According to Ragge (1965) this species prefers dry habitats and appears to thrive in areas where there are bare patches of ground such as waste-land. It is rarely found where there is lush growth. Neverthe- less, C.brunneus is found in a wide range of habitats ranging from meadows to the sides of slag tips. Probably no other British grasshopper can exist in such extremes of environment and for this reason C. brunneus is ideal material in which to study differences in individuals living in various habitats.

There has been no previous quantitative work on the ecological distribution of the different morphs of C. brunneus. Neither is there much information about the inheritance of the different morphs. Sansome & La Cour (1935) demonstrated the presence of at least fourteen genes which showed epistatic inter- actions in Chorthippus parallelus (Zetterstedt). Harvey ( I 977) showed the inheritance of discrete colour patterns in Myrmeleoterrix maculatus (Thunberg). Nabours ( I 929) and Nabours et al. J1932) demonstrated at least

Correspondence: Dr Peter D. Gill, Department of Genetics, The University, Nottingham.

twenty-four dominant genes for patterns in grouse locusts (Tettigidae).

Mcthods

Collection

Grasshoppers were collected from twenty- two different sites in Nottinghamshire, Derby- shire and Essex (Table 1 ) using a small net; the collection area did not exceed 50m diameter. After collection, the insects were kept in the laboratory according to the methods of Kelly-Stebbings & Hewitt (1972).

Scoring

Only sexually mature adults were scored in this survey because colour changes occur in the nymphal stages (Richards & Waloff, 1954) usually between instars. One individual changed from light green to dark green in the final instar but before sexual maturity had been reached.

Colours

Colours were scored as green, purple, white, black, or brown (Table 2). The latter category included light-greys, dark-greys, light and

03074946/79/08004249 $02.00 0 1979 Blackwell Scientific Publications 249

250 PeterD. Gill

TABLE 1. Locality related to the frequencies (%) of different wing-types and the wedgeless character ~ ~~ ~

Wing type Locality 0,s. reference Wedgeless Sample

Striped Stripamottled Mottled Plain size

Habitat type I Epping Forest 1 Epping Forest 3 Frinton 2 Frinton 3 Kingston Lockington Shipley

Habitat type 2 Dovedale Epping Forest 2 Frinton 1 Kimberley Mansfield 2 Nottingham Selrton

Habitat type 3 Ilkeston Jaywick Kirkby Mansfield I Point Clear Staple ford St Osyth Woodford Green

T Q 913396 TQ 967417 TM 210251 TM 249240 SK 271490 SK 201473 SK 451432

SK 513140 TQ 905394 TM 208232 SK 440494 SK 552565 SK 452532 SK 520475

S K 4 13464 TM 140153 SK 547405 SK 560539 TM 140093 SK 389409 TM 117129 T Q 9074 10

24.4 21.3 17.6 25.5 25.6 33.3

5.7

31.4 20.9 21.7 20.6 11.1 36.1 13.0

21.2 25 10.3 2.7

19.6 0.6

14.3 2 0.3

14.6 36.6 10.2 34.1 17.6 29.4 17.0 29.0 20.5 35.8 16.1 33.3 17.1 40.0

11.4 40 7.0 53.5

17.4 45.7 17.6 41.2

5.6 61.1 13.3 23.3 13.0 52.2

15.2 57.6 9.6 50.0 3.4 75.9 0. 1 75.7

30.4 37 8.6 51.4

22.9 40.0 0. I 47.0

24.4 20.4 35.3 17.1 11.9 16.7 37.1

17.1 10.6 15.2 23.5 22.2 26.7 21.7

6.1 15.4 10.3 13.5 13 31.4 22.9 15.2

9.0 20.4 11.8 10.6 0.0 2.0 5.7

14.2 23.3 21.7

5.9 11.1 10.0 17.4

12.1 11.5 10.3 16.2 10.9 20.0 17.1 19.6

4 1 44 5 1 47 39 36 35

35 44 4 6 34 18 30 23

35 52 29 37 4 6 35 35 46

TABLE 2. Numbers of different coloured grasshoppers related to locality

Locality Purple Green Black White Brown

Habitat type I Epping Forest I 4 0 2 2 33 Epping Forest 3 0 0 3 3 38 Frinton 2 I 2 1 1 46 Frinton 3 3 I I 3 39 Kingston I I 0 2 35 Lockington 0 0 0 2 34 Shipley 4 0 I 0 23

Habitat type 2 Dovedale I 0 I 2 31 Epping Forest 2 I I 1 4 37 Frinton 1 1 2 4 1 38 Kimbcrley 0 2 0 2 30 Mansfield 2 2 1 0 0 1 5 Nottingham 0 0 0 3 27 Selston 0 0 0 3 20

Habitat type 3 nkeston 2 0 2 0 3 1 Jaywick 3 0 0 6 43 Klrkby 0 0 0 0 29 Mansfield I 0 I 1 0 35 Point Clear 1 1 2 1 41 Staple ford 2 0 I 1 31 St Osyth 0 0 5 7 23 Woodford Green 3 I 2 0 40

Colour patterns in Chorthippus 25 1

Characters of the head and pronoturn

Discrete characters were scored (Fig. 1). Most are described by Ragge (1965) and are as follows: (i) Presence of linea intermedia. (ii) Presence of fascia postocularis. (iii) Presence of linea media or h a media. (iv) Presence of the wedge. This term is used differently to that of Ragge ( 1 965). In this paper the wedge refers to the carina lateralis; when absent, the fascia dorsalis interna and fascia dorsolateralis are either faint or not observed. (v) Presence of zona lateralis.

dark browns, and greenqeys. The range of different colours in the last category is great and there appears to be considerable overlap. Using a colour atlas (Kornerup & Wanscher, 1963) the various colours were identified as follows:

Brown: 4B2-5; 4C2-3; 4D3-4; 4E3-5; 4F6-8; 5C2-8; 5D2-8; SE2-8; SF4-8; 6B2-4; 6C2-5; 6D6-8.

Green: 2785-6; 2887-8; 29D7-8. PurpIe: 12C3-4; 12D4-5; 12F7-8. White: 4A2;4A3. Black: 5F2; 5F3.

Because this was a preliminary investi- gation, and in the absence of breeding data, the various shades of a colour were combined. Colour was scored from the dorsal pronoturn.

Lino intermedia Lima loteralis

k i n a lateralis

W w pottern

Zono loterolis Lineo scopularis Stigma

Wedge

FIG. 1 . Diagram, rhowing charactera o f the head, prothorax and forewings (modified from Ragge, 1965).

Forewing types

Forewings were classified according to four types: mottled, striped, stripe-mottled or plain. The striped and mottled varieties are shown in Fig. 1. The stripe-mottled variety has both characters whereas the plain variety has no pattern on the forewing.

Other forewing characters scored

(i) Presence of linea scapularis: Ragge (1965) notes that this character is rarely found in males. (ii) Presence of the stigma.

Characteristics of the hind femur

(i) The upper surface of the hind femur is usually lighter coloured than the outer lateral side, which may be heavily pigmented. This makes the background colour of the upper

Barred Mottled Barred-mottled Plain

FIG. 2. Diagrams showing different characters of the hind femora.

252 Peter D. Gill

0 0 1 I I I I

surface of the femur stand out. (ii) Nearly all C.brunneus were found to have bars on the upper surface of the hind femur (Fig. 2a) although the intensity of pigmentation varied enormously. (iii) The presence of light mottling on the upper surface of the femur was recorded only if distinct (Fig. 2b) since the presence of bars on the femur tended to mask the light- mottling effect completely. Where both characters were obvious in the same individual, it was recorded as having barred-mottled femora (Fig. 2c). (iv) If no pattern was dis- tinguishable on the upper surface of the femur, it was recorded as plain (Fig. 2d).

Classification of collection sites

After a collection, all common species of grass and flowering plants were noted and the area was designated one of three broad ‘types’ of habitat (Table 1) namely:

Type I : ‘Uniform’ habitats. These are areas which had little bare ground and were either green, e.g. Shipley, or meadows with long grass, e.g. Epping Forest 1. Type 1 habitats were well covered with foliage. Only a few dominant species of plant were present.

Type 2: Intermediate between types 1 and 3, including well-covered wasteground with patches of bare ground. A number of different species of grasses, flowering plants and bushes were present.

Type 3: Large areas of bare ground con- taining patches of vegetation. Sites included the sides of a slag tip (Kirkby), side of a sea wall (St Osyth), land adjacent to coniferous woodland, in an area covered with patches of long grass and bracken (Mansfield I ) , and sparsely covered waste-ground (e.g. Woodford Green, Point Clear, Jaywick, Ilkeston).

A nalysis of data

Data from the three different habitat types were compared with each other using three- dimensional contingency tables. The pro- cedure followed is described by Fienberg ( 1 970) and Everitt ( 1 977). All calculations were made using the Generalized Linear Interactive Modelling program on the com- puter at Nottingham University. The significance of associations between two

characters was determined by examining standardized values (Table 3) of the calculated parameter estimates (Everitt, 1977, pp. 94-99). Standardized values have a standard normal distribution and are compared with the 5% normal deviate for significance. Table 3 shows the percentage of a particular forewing type having a particular character (for example, 88.1% of stripe- winged individuals in habitat type 1 have a zona lateralis). x 2 testing for associations was carried out on the relevant frequencies, however.

Wing pattern frequencies were compared using t or d tests (Bailey, 19S9, pp. 47-51) after applying an angular transformation to the data. Wing pattern frequencies are shown in Table 1 .

Results

Wing patterns

The frequencies of striped related to mottling in different areas are shown in Fig. 3. The mottling character predominated in type 3 habitats (P< 0.01 between types 1 and 3; t-test). Conversely, the striped character was commonest in areas of long grass land ( P < 0.05; d-test between types 1 and 3). Because Shipley was an atypical habitat with little long grass and much lush short grass and areas of clover, it was omitted from the analysis. The results suggest that the mottling character may be selected for in sparsely covered habitats, whereas the striped character may offer

40cx 0

0 X X 0

TA

BE

E 3

. A

ssoc

iatio

n of

cha

ract

ers

of t

he p

rono

tum

and

hea

d w

ith

thos

e of

the

fore

win

g (96).

The

dat

a fo

r th

e fa

acia

pos

tocu

luia

wer

e ca

lcul

ated

on

graa

ahop

pers

po

sses

sing

a li

ner

inte

rmed

ia. S

igni

fica

nces

of s

tand

ardi

zed

valu

es o

f pa

ram

eter

est

imat

es w

e re

cord

ed (*P< 0.

05; **P< 0.

02; .**P<0.01; *

***P

<0.

001)

.

Stan

dard

ized

val

ues

of

para

met

er e

atlm

ates

(w

ing

char

acte

r W

ing

char

acte

r D

egre

es

com

pare

d w

ith

pron

otum

or

fem

ur c

hara

cter

) H

abita

t

type

St

ripe

d St

ripe

- M

ottle

d Pl

ain

free

dom

St

ripe

d St

ripe

- M

ottle

d Pl

ain

X'

of

Cha

ract

er

mot

tled

m

ottl

ed

Zon

a la

tera

lis

Une

a m

edia

/ ca

rins

med

ia

Lin

er

inte

rmed

ia

Firc

ia

port

ocul

aris

Wed

peie

ss

Dar

k ou

ter

uppe

r su

rfac

e of

hin

d fe

mur

1 1 3 I 1 3 I 7 3 I 1 3 I 1 3 I 2 3

88.1

94

.3

90.7

77.6

71

.7

64.8

88.1

73

.6

74.1

30.5

30

.8

32.5

0 0 2.0

1.5

5.7

1.85

86.2

89

.7

85.7

70.6

55

.1

69.0

68.6

82

.8

76.7

68.6

37

.5

63.6

0 1 .o

0 1.

96

6.9

2.3

20.2

19

.4

18.7

10.1

7.

8 3.

6

16.2

13

.6

16.3

31.3

21

.4

7.4

20.2

31

.1

24.7

13.1

23

.3

16.9

31.2

86

2.0

26.4

46

.0

6.6

2024

.0

7.5

2.0

46.1

80

5.7

49.1

48

.0

77.1

23

1.3

69.2

79

.2

7.9

722.

3 3.

8 8.

0

18.4

11

88.0

28

.3

20.0

8 5

.6**

**

3.5

****

8 1.

95

1.00

8 4

.14

****

2.

48..

8 -3

.95

****

1.

21

8 -7

.39

0

* 4

-1

4

8 -5

.49 .. . .

-3.8

5 .**

-2.3.

-4.0

8**.

*

-0.3

7

3.2

6**

*

-6.6

6***

.

-3.3

8***

*

254 PeterD. Gill

better camouflage against long grass. No significant differences were found for the stripe-mottled and plain characters, although if the Shipley data was reintroduced into the analysis a positive association between type 1 habitats and the plain character was observed (P < 0.05; t-test between types 1 and 3).

The linea scapularis was found in virtually all females except for a few mottle-winged- wedgeless ones. Similarly, nearly all male and female grasshoppers had a stigma, although it was absent in a few mottle-winged-wedgeless individuals.

Forewing characters related to those of the pronotum

Data were analysed as three-dimensional tables (Table 3). In all cases examined significant associations were recorded between character, wing pattern and habitat type.

Zona lateralis. Table 3 shows that there is a significant positive association between the striped and stripe-mottled varieties (P < 0.00 1) whereas mottled and plain-winged were negatively associated with this character (P < 0.001 and < 0.05 respectively).

Linea medialcenrra media. A negative association was recorded for the mottled and plain characters with the linea media and canna media (P< 0.001, Table 3). Grass- hoppers in habitat type 1 are more likely to have a linea media and carina media than in habitat type 3 (P < 0.05 ; 2 x 2 x2 test).

Linea intermedia. A positive association between striped and stripe-mottled grasshoppers with the linea intermedia was recorded (P < 0.001 and < 0.02 respectively, Table 3). The mottled variety showed a negative association with this character (P < 0.001). Area 1 had significantly more striped grasshoppers with the linea intermedia compared to those from habitat type 3 (P< 0.0s; 2 x 2 xz test).

Fascia postocularis. This character was recorded as the presence of a coloured (usually grey) area between the linea lateralis and linea intermedia (Fig. 1). Hence the data in Table 3 were calculated only on individuals possessing a linea intermedia. Striped and mottled individuals were negatively associated with the fascia postocularis (P < 0.001 and < 0.01, Table 3) whereas plain individuals

were positively associated (P < 0.0 1). There were fewer than expected stripe-mottled indigduals with a fascia postocularis in habitat type 2 (P< 0.05; 2 x 3x2 test).

Wedge. There is a positive association between wedging and all forms of forewing pattern (P < 0.00 1, Table 3), which indicates the comparatively high frequency of this character. More mottled and plain individuals than striped and stripe-mottled individuals do not have a wedge (P < 0.001 ; 2 x 2 x 2 test).

Wing pattern and the upper surface of the hind femur

Table 4 demonstrates that a dark femoral upper surface was uncommon but occurred more often in the plain-winged variety than in other varieties (P < 0.01 ; 2 x 2 x 2 test). More dark femoral upper surfaces in mottled and plain-winged individuals were observed in habitat type 2 than in types 1 and 3 (P< O.OOI;~X 4 x 2 test).

A positive association was observed between barred-mottled femora and mottle- winged individuals (P < 0.02, standardized value of parameter estimate, Table 4). The proportion of barred-mottled femora in mottle-winged individuals increases from 40.7% in habitat type 1 to 66% in habitat type 3 (P< 0.01 ; 2 x 3 x 2 test; Table 4). A positive association was also recorded between mottle-wings and barred femora (P < 0.00 1 , standardized value of parameter estimate).

Colour variation

Data were treated as a 5 x 4 X 3 contingency table in order to examine colour variation in relation to forewing type and habitat type. After omitting the results for brown grasshoppers, the data were reanalysed as a 4 x 4 x 3 contingency table. By examining standardized values (Table 5 ) unequal frequencies of wing type within each colour class could be determined.

Purple. The rarest variety was that with purple dorsal pronotum/purple sides. One was found at Stapleford and another at Ilkeston. Purple individuals were found in several populations (Table 2). however. At the Woodford Green site there were many dead dark purple-brown dock (Rumex

Colour patterns in Chorthippus 255

TABLE 4. Contingency table showing the numben of hind femur characten In relation t o wing characters and habitat type (x' = 1610 for 24 degrees of freedom)

Hind femur character f i b i t a t wing type character b e d B . m d - Mottled Plain

mottled

2

1 Striped Stripe-mottled Mottled Plain

Striped Stripe-mottled Mottled Plain

3 Striped Stripe-mottled Mottled Plain

46 37 7 0 53

45 24 02 44

45 40

122 30

14 7

22 I t

3 4

15 6

6 3

35 9

7 7 5 8

4 I 4 2

0 0 5 2

0 0 0 4

1 0 2 1

0 0 2 2

obfusifolius) plants. Grasshoppers from this area were noticeably darker (12F7-8 in the colour atlas; Kornerup & Wanscher, 1963) than those from an area such as Frinton 2 where the purple in the environment was provided by the flowers of thistles. Individuals were classified as colour 32C3-4. Purple shades occur commonly in different habitats and may be associated with purple grau- hoppers. At Shipley, purple in the environment was provided by clover (TrijWiurn pratense) at Mansfield 2 by dead bracken (Aquilinum sp.), by bramble (Rubus

frutiosus) at Jaywick and by marjoram (Origanum vulgare) a t Dovedale.

Most grasshoppers observed were purple plain-winged (twenty-two plain, three mottled, two striped; P < 0.001, Table 5b). Wings were either light brown or purple. Significantly fewer purple grasshoppers were observed in habitat type 2 (P< 0.05; 2 x 2 x z test between habitat types 1 and 2) .

Green. Twenty-one out of twenty-four green grasshoppers observed had either plain light brown or light green wings; the remaining three had mottled wings (P < 0.001, Table

TABLE 5. Standardized parameter values for habitat type and wing pattern. (a) Raul t s from a 5 X 4 X 3 contingency table. (b) Results from a 4 X 4 X 3 contingency table in which brown grasshoppen were omitted. R r a m e t e n not recorded were not significant.

X' d.f. Parameter value

Parameter

(a) Habitat 1, brown Habitat 2, brown Habitat 3, brown White, plain Brown, stripe-mottled Brown, mottled Brown, plain

(b) Purple, plain Green, plain Black, mottled Black, plain White. plain

7.6 1 1.29

2.4 7

10.52

8.44

-2.82

-4.58

4.32 3.95 2.25 1.31 1.89

< 0.001 < 0.00 I < O . O O l . < 0.02 < 0.01 < 0.00 I < 0.001

< 0.00 I < 0.00 I < 0.05 n.r. < 0.001

1450 30

113.3 24

256 PeterD. Gill

5b). The greatest frequency of green grasshoppers was observed at Shipley (Table 2). This habitat was an unusual place to find C. brunneus in that it consisted of large areas of green clover and short grass. It was also very damp. Possibly a greenstriped wing pattern as found in M.macuhtus (Ragge, 1965; Harvey, 1977) would be conspicuous in the absence of long grass.

Green C.brunneus were found to be commonest in habitat type 1 (P < 0.01 ; 2 x 2 x2 test between habitat types 1 and 3).

White. In the white variety, the wings are always white and usually plain (forty-three plain, one stripe-mottled, one striped, one mottled; P < 0.001, Table 5b). There is usually a prominently dark fascia postocularis present. The presence of the white variety at St Osyth (Table 2) may be explained by the large areas of light grey concrete upon which grasshoppers frequently basked; also the white coloration was a darker shade than that found elsewhere (intermediate shade between colours 5A2 and 5B2 in the colour atlas; Kornerup & Wanscher, 1963). No significant differences between the white variety and habitat type were observed. Possibly it evolved as camouflage against the dead stems of grasses.

Black. The black variety was predomi- nantly mottled or plain (fourteen mottled, eight plain, two stripe-mottled, two striped). Table 5(b) shows a positive association between black and mottled forewings (P < 0.05). It occurred at low frequency in most habitats (Table 2); there were no significant differences between habitat types, however. Harvey (1977) noted that the black variety of M.maculatus was common on recently burnt ground. The black variety of C. brunneus was common at St Osyth possibly because of the presence of black tar between the stones of the sea wall. A black grasshopper on the light grey concrete may be conspicuous to a predator, however.

Brown. The brown coloration was pre- dominant in all three habitat types (P < 0.001, Table 5a) and extremely variable, being striped ( I66), stripe-mottled ( 1 20), mottled (347) or plain (85). Brown and mottled were positively associated (P < 0.001) whereas stripe-mottled and plain were negatively associated (P < 0.00 1). The shades of brown

may be affected (either genetically or pheno- typically) by the general colour of the environment; for example, at Kirkby the pre- dominant colour was a light grey which blended well with the background colour of the slag tip. The predominance of the brown variety probably reflected the preference of this species for fairly bare ground.

Discussion

Experiments by Richards & Waloff (1 954, p. 54) on M.macuIatus did not give conclusive evidence that environmentally induced changes in coloration of the grasshopper could occur. Ragge (1 965) noted that certain African grasshoppers could increase their black pigments on dark backgrounds. Harvey’s (1 977) breeding experiments on M . maculatus suggest that the colour characters segregate discretely and that colours did not change from nymph to adult. Intensity was not measured, however, so some darkening or lightening may occur.

Sansome & La Cour (1935) isolated fourteen genes and several others were indi- cated in C.parallelus. Six of these genes inhibited the expression of at least two non- allelomorphs. One phenotype may therefore correspond to a comparatively large number of genotypes and by the presence of this epistacy the species is able to react readily to different habitats while remaining fairly uniform in one habitat.

Nabours ( 1 929) and Nabours et al. ( 1 932) showed that genes for twenty-four co- dominant colour patterns exist in the grouse locusts Acrydum arenosum (Burmeister) and Paraterfix texacanus (Hancock). Linkage was found to occur but the linkage relationships differed significantly between the two species (in C.paraflelus linkage was not well marked).

The results of the present survey suggest that if the colour patterns of Cbrunneus are genic, the association between colour and wing type may be due to selection producing gametic phase imbalance, which could be aided by linkage. Grasshoppers are known to be preyed upon by birds and lizards (Ragge, 1965). Patterns which were found to be associated with each other or with certain colours are as follows: (i) Plain wings were positively associated with a dark upper surface of the

Colour patterns in Chorthippus 257

Dr D. T. Parkin for statistical advice. This work was supported by an SRC post-doctoral research fellowship.

femur, a dark fascia postocularis and a purple, white or green pronotum. (ii) The brown pronotum was positively associated with mottled forewings. (iii) Striped forewings were positively associated with the linea intermedia, linea media/carina media, zona lateralis and a light upper surface of the hind femur. (iv) Individuals with mottled forewings were more likely to be wedgeless and were positively associated with barred-mottled hind femora.

Associations with habitat type A significantly higher frequency of brown

stripe winged grasshoppers were found in habitat type 1. Proportionally more of these individuals had a linea intermedia and a linea media/carina media than striped individuals from other habitat types. In type 3 habitats a significantly higher frequency of mottled grasshoppers was observed.

In conclusion, longitudinal markings on the pronotum and forewings often occurred together in the same individuals and a positive association with habitat type 1 was observed. Possibly these characters provide camouflage in long grass. Conversely, the greater amount of bare ground found in type 3 habitats may favour mottle-winged individuals in which longitudinal markings on the pronotum are absent.

Acknowledgments I am grateful to Dr C. M. Hewitt for his useful comments on the manuscript and to

References

Bailey. N.T.J. (1959) Statistical Methods In Blology. English Universities Re=. London.

Everitt. B.S. (1977) The Annlysis of Condngmcy Tab&;. C h q m m & H.11, London.

Fmnberg, S.E. (1970) The analysis of multidimen- sional contingency tables. €cology. 5 1 . 4 1 9 4 3 3 .

Harvey, A.W. (1 977) The expruaion of supernumary chromoromes in Mytme&otettfx rnocubhrs (Thunbcrg) (Acridse). Unpublished Ph.D. thesis, University of E u t Anglia.

Kelly-Stebbing. A.F. & Hewitt. G.M. (1972) T h e laboratory breeding of British Gomphocerinc grasshoppen (Acrididre: Orthoptera). Acridu.

Kornerup. A. k Wanscher. J.H. (1963) Methuen Handbook of Colour. Methuen. London.

Nabours. R.K (1929) T h e genetics of the Tetti- gidae (grouse locusts). Bibliogmphh Ccneticu. 5 .

Naboun, R.K.. Lamn. 1. & Hutwig. N. (1932) Inheritance of colour patterns in the grouse locust Acrydurn urnnonun Burmeister (Tetti- gidae). Genetics, 18. 159-171.

Ragge. D.R. (1965) Grasshoppers. Crfckets and Cockroaches of the British Isles. Frederick Warne. London.

Richards, O.W. & Wdoff, N. (1954) Studies of the biology and population dynamics of British grasshoppers. And-locust Bulktin. 17. 1-1 82.

Sansome, F.W. & La Cour. L. (1935) The genetics of grasshoppers: Chorthippus purulklus. Journal of Genetics, 30.415-422.

1.232-245.

27 -1 04.

Received 5 December 1978

18

Colour-pattern variation in relation to habitat in the grasshopper Chorthippus brunneus (Thunberg)

Documents