Table 2.

Regression models for factors affecting: (A) mass, (B) wing loading, (C) wing length and (D) wing beat frequency in D. melanogaster

Variable*Parameter estimates.e.m.t-valueP-value
A. In mass (mg):
F4,277=438, P<0.001, R2=0.86Intercept0.41720.0277815.02<0.001
Tdev (°C)–0.00820.00112–7.31<0.001
T dev 2 –0.00100.00022–4.58<0.001
Gender (males)–0.19590.03441–5.69<0.001
Tdev × gender–0.00570.00152–3.76<0.001
F3,254=530, P<0.001, R2=0.86Intercept1.56450.0664623.54<0.001
Tdev (°C)0.07070.0021033.70<0.001
In mass (mg)1.96220.1275515.38<0.001
Gender (males)0.26230.044815.86<0.001
C. Wing length (mm):
F3,225=680, P<0.001, R2=0.90Intercept1.79180.1521811.77<0.001
Tdev (°C)–0.00880.00241–3.64<0.001
Wing area (mm2)0.23260.029187.97<0.001
Gender (males)–0.10510.02052–5.12<0.001
D. Wing beat frequency (s–1):
F5,102=86.23, P<0.001, R2=0.81Intercept142.406618.33807.77<0.001
Tdev (°C) 15°C–8.16273.4674–2.350.020
Ttest (°C)7.75250.590213.14<0.001
In mass (mg)55.937013.91074.02<0.001
Wing area (mm2)–24.54104.8130–5.10<0.001
Gender3.67413.72590.990.326
• F, P, and R2 regression statistics are provided for the overall model, followed by the parameter estimates, standard error, t-value, and P-value for each variable within the model

• * If parameter estimates are used for prediction purposes, the following should be noted: variables with squared terms are centered on the variable's mean, consequently, Tdev2 equals (Tdev–22.1)2. When estimates are calculated for males the parameter estimate for gender is included in the model, for females the parameter estimate becomes 0. Because the regression estimates for Tdev are based on only three temperatures, caution should be exercised if these models are used for prediction at developmental temperatures other than 15, 23, or 28°C

• Includes flies from both the flight' and lift' categories of performance, but not failed' flight. Flies that performed a flight' or generated `lift' are grouped in the analysis because there was no significant difference in wing-beat frequency (WBF) between these flies (when flight performance was included as a variable in the above analysis, P=0.88). Flies that developed at 23°C and 28°C were grouped because WBF differed by only 1.3 beats per second as estimated by the model and was not significantly different