The question cannot be answered. You need to know the infection rate in the vector. If 0.0001% of the sandflies are carrying the disease you will need a sample size in the hundreds of thousands. Likewise if 99.999% of the sandflies are carriers. If only half of the females are carriers, then your sample size can be quite reasonable.

If the infection rate is less than about 20%, you should consider pooling your samples. To do this you must first figure out the sensitivity of your assay. If I have 1 infected female in a sample of 2, 4, 8, 16, or 32 non-infected females can I still detect the infection? As infection rates get smaller and smaller, you can increase pool size up to the limit of reliable detection. Eventually you will just have to say that the infection rate is so low that available resources were insufficient to detect if the pathogen was present or not.

If the infection rate is higher than 20%, say 99.9% then you will just have to process a large number of samples to get the level of accuracy you require. Pooling will not help.

See my 2010 article "Reexamining the Pooled Sampling Approach for Estimating Prevalence of Infected Insect Vectors" (attached).

i agree with Timothy A Ebert

It is best to conduct a survey of vector insects and establish positive infection rates and then conduct molecular examination of positive samples only. May be useful to look to this survey:

Article Isolation and Identification of some blood parasites from mi...

With all best wishes

Afkar