I think that it is not a problem of data transformation, but of data interpretation.

You can simply transform negative values in positive ones and do arcsine or log transformation. The problem you have it is probably to explain why you found more population after the treatment.

Good luck with your research!

Thank you very much.

Dear Ebru, Several issues should  be considered:

1.  He-Ti formula is designed and works  with  experimental  data   with precondition that  pesticide treated area  will be  more   successful than control.  That is  why in some cases when we have   such  kind unusual situation:  Treated area  is less effective than  the control, this  Formula gives for sure negative results.  But  this is interpretation which you are looking for...

2. I don't think that any type of  statistical  transformation of data  will be useful  and changes anything in case of interpretation of results - I agree with Paolo...

3. You can  use   ANOVA in  regular table, make transformation of  percentage data   in several way.  I using SPSS, where you can    transform  experimental data by  option compute variables...   i think logarithmic transformation is better solution..

4. There been in Research gate   interesting discussing  on similar issues: lhttps://www.researchgate.net/post/How_can_I_apply_an_arcsine_transformation_to_my_data_using_Statistica - might will be helps you 

5. This is very interesting results -  control has better achievements that treated - it will be quite difficult  to explain this...

6. What  is  Latin name of white grubs you are worked on  ? We have  grubs -  (Hypoaspis) Polyphyllae olivieri, which is extremally difficult to control...Only way use high rates  nitrogen fertilizers  incorporate in soil several  times...

Thank you for your attention. It was a biopesticide experiment on Melolontha spp.