We investigated the responses of cotton (Gossypium hirsutum L.) to the
combined effects of soil water deficit and twospotted spider mite (Te
tranychus urticae Koch) infestation. Two mite treatments (-M: uninfest
ed, +M: artificially infested 83 d after sowing), and two water regime
s (+W: well watered, -W: water stressed) were combined factorially in
four treatments. Mite colonies developed at similar rates in well-wate
red and water-stressed crops. Despite the similar intensity of infesta
tion, visual symptoms of mite injury were more marked in well-watered
host plants (+M+W) than in their water-stressed counterparts (+M-W). L
int yield of unstressed controls (-M+W) was 175 g m(-2). In uninfested
crops, water deficit reduced yield by 30%, mites reduced the yield of
well-watered crops by 92%, and the combination of mite infestation an
d water deficit reduced yield by 72% (water effect: P < 0.01; mite and
interaction effect: P < 0.0001). Differences in yield responses to mi
tes between well-watered and water-stressed crops were mostly related
to differences in reproductive partitioning. The interaction between m
ites and water deficit was also significant for other crop variables i
ncluding canopy temperature, leaf water potential, concentration of ni
trogen in reproductive structures and seed oil concentration. The magn
itude and consistency of the interaction between both stresses indicat
es that, under our experimental conditions, mechanisms of adjustment t
o water deficit may have enhanced cotton resistance to miles. This is
further supported by (a) an increase in specific leaf weight and a par
allel increase in leaf penetration resistance due to water deficit; (b
) a negative association between macroscopic symptoms of mite injury a
nd leaf penetration resistance; and (c) a choice rest showing that adu
lt female mites preferred to feed and oviposit on leaves from well-wat
ered plants. (C) 1998 Annals of Botany Company.