I am not sure if my guess is correct.
As you mentioned GENMOD with logit link and binomial distribution, I guess the dependent variable is on which leaf the aphid stays. I name the dependent variable as BYDV-Infected, by assigning BYDV_Infected=1 if the aphid is on the leave of BYDV-Infected, and BYDV_Infected=0 if the aphid is on the leaf that was uninfected.
800 aphids are used in this experiment, and every 20 minutes a record is made to write down if the aphid stays on the infected leaf or on the uninfected leaf. This implies the existence of repeated measurements. Hence I suppose each aphid has a unique identification number which is aphid_id.
Let’s suppose the independent variables are age of the aphid, sex of the aphid, size of the aphid and species of the aphid, and the name of the data is bioassay.
The SAS code in this situation for a GEE1 model is:
proc genmod data=bioassay descending;
class species sex;
model BYDV_Infected = age sex size species / link=logit dist=bin type3;
output out=out p=p;
repeated subject= aphid_id /type=cs;
run;