Female tsetse flies, Glossina pallidipes Austen, caught in artificial refug
es were subjected to ovarian dissection and analysed for levels of fat, res
idual dry weight (RDW) and haematin. There were rather small proportions of
flies in ovarian categories 0 and 1, in part due to large losses in the im
mature and teneral stages at the hottest time of year. The distribution of
the female catch among pregnancy days was close to uniform. The wet and dry
weights (WW and DW) and RDW of eggs, larvae and pupae increased by 0.821,
0.303 and 0.204 mg respectively, with each mm(3) increase in volume. Water
accounted for 71.7% of the fat-free WW, and fat for 32.7% of the DW. Betwee
n birth and ovulation, fat increased from 2 to 4 mg and RDW from 7 to 11 mg
; thoracic RDW increased by 2.5 mg and changed little thereafter. Fat level
s increased 3.5 mg by day 6 of pregnancy, but only 0.5 mg thereafter. Over
the same periods RDW corrected to zero haematin (CRDW) increased by 1 and 8
mg respectively. Full-term fat and CRDW levels were 8.2 and 19.3 mg respec
tively. Cumulative haematin frequencies formed a smooth curve with a slope
that increased continuously. The raw data were well fitted by a model where
feeding rates increased exponentially and capture probability was independ
ent of haematin content. The mean feeding interval was 60 h; feeding probab
ilities of > 0.9/day were only found in flies that had failed to feed for >
72 h. in early pregnancy, fat levels declined with haematin for flies that
had fed > 36 h previously; by days 5-7 fat levels were maintained at a con
stant high level for 60 h post-feeding. Fat-haematin graphs for female tset
se cannot be used to estimate rates of fat utilization. Traps sample tsetse
with below-average fat and RDW in early and late pregnancy respectively. R
efuge samples are less biased than those from traps; they give a better pic
ture of the dynamics of pregnancy in normal flies and facilitate the explan
ation of existing anomalies.