1. The work presented in this thesis is concerned with the study of
stabilising selection for body size in D.melanogaster.
2. The method followed in this study was to create differences in body
size through selection and observe the changes in fitness associated with
these differences. Also, once the differences were created, we observed
the effects of natural selection on the selected lines when selection
3. The consideration of this problem involves the choice of a suitable
estimate of fitness as well as the definition of a suitable ecological
system in which fitness should be measured. We decided that the best
indicator of fitness available is the biomass produced per unit of time
in competition with a standard competitor, in a live -yeast system.
4. The physiological changes responsible for the changes in body size
may or may not affect fitness, tinder the conditions described. Given
hypothetical differences in physiological parameters like feeding rate,
efficiency of food conversion and critical size we were able to simulate
in a computer the outcome of competition between strains differing in
one or more of these parameters. This allowed us to make predictions of
the changes in competitive ability and the productivity of the cultures.
Therefore we were able to analyse our results in terms of a system
rather than in terms of individual characters.
5. It was found that the physiological changes responsible for the changes
in body size of the selected lines were probably changes in feeding rate
associated with changes in the critical size. However, the "nature" of
growth is affected in this case since we can detect differences in body
size in flies exhibiting the same development time. This change in the
"nature" of growth can be explained if we postulate an unbalance of growth
in the selected lines.
6. Although the changes iñ body size obtained by selection are quite
sùbstantial, it was difficult to detect a consistent change in compeitive
ability measured against a standard competitor. It is suggested that this
difficulty might be due to changes in the variance of the growth parameters
correlated with changes in their mean. Within the range of deviations
of body size from the unselected value these changes in variance would
counteract the possible changes in fitness which are a consequence
7. Relaxation of selection for lines selected for large and small body
size had little effect in bringing the mean fialue of the character back
to the unselected level. One possible exception was verified when selection
was relaxed in a population cage.
Back selection had an immediate response, similar to that of forward
selection. The experiments which tested the competitive ability of the back
selected lines were not conclusive. Selection for short development time
in the selected lines did not affect body size;,the changes in competitive
ability were not well defined.
Selection for fast and slow development had some response when
development time was measured under pure culture conditions. Under
competitive conditions the apparent advantage of the fast line disappeared
but the disadvantage of the slow line persisted. The response to
selection for fast and slow development was accompanied by a reduction
in body size below the unselected level in the fast line and an increase
in the slow line above the same level.
8. Selection for large and small body size under competitive conditions
showed some response in both directions though less well marked than when
selection was carried out under optimal conditions.
9. Inbreeding caused a proportionally equal decrease in body size in all
the lines. This decrease was more accentuated when body size was measured
under competitive conditions. Competitive ability was affected differentially
in the 1st generation of inbreeding, but this difference disappeared
subsequently. Viability was reduced below the non -inbred level.
10. The experiments on egg production of the selected and unselected flies
grown under different conditions in the larval stage fed different amounts of
food in the adult stage revealed a superiority of the unselected over the
11. The crosses between the Pacific and the Kaduna populations showed no
breakdown or improvement in competitive ability in the Fl or the F2,
suggesting that the genes controlling the growth ability which is correlated
with competitive ability are the same in the two populations.