Retention of memory through metamorphosis: can a moth remember what it learned as a caterpillar?

Three days post-eclosion, individual adult M. sexta of each treatment were placed in the loading arm of the Y choice apparatus described above. Again, ambient air or EA-impregnated air was pulled through separate arms of the apparatus to the central chamber, and adults were allowed to walk between the arms for 10 minutes. The entire test was conducted in the dark. Upon completion of the 10 minute testing period, the location of the adult was scored by a blind observer as either the EA arm, ambient air arm, or no choice. The entire apparatus was rotated 180° between batches of individuals.

To add EA contamination, we applied 10 µl of a 2% agarose gel impregnated with a 10% ethyl acetate solution to the dorsal thoratic region of untreated control pupae. Gel was used to ensure that no liquid EA entered the spiracles, and to decrease the volatilization rate of the odor. Agarose was refreshed twice weekly, starting three weeks after pupation for a total of four to six applications. To remove any possible external contamination, the pupal cases of trained larvae (shock+odor) were washed thoroughly with a soft brush and distilled water one and three weeks after pupation (methods from Barron and Corbet, 1999). Three days post-eclosion, individuals from both treatments were tested for odor aversion in the Y choice apparatus, described above. In addition, untreated control larvae were washed to determine if mechanical stimulation at the pupal stage resulted in in a change of adult behavior.

All statistical analyses were performed using SPSS v. 14.0 (SPSS Inc., Chicago, IL). For each treatment the proportion of larvae or adults choosing the ambient air arm vs the EA arm was compared to an expected value of 50 percent with a two-tailed binomial calculation. Two-tailed tests were employed in order to detect a possible attraction to EA in contamination experiments. Constancy, the percentage of adults that made the same choice in the Y chamber as they did as larvae, was examined with a two- tailed binomial calculation with the expectation that fifty percent of individuals choosing no odor as adults made the same choice as larvae. Learning between larvae trained at third and fifth instar was compared with a chi-squared test of equality of distributions. Power tests were conducted with α = 0.05 and β = 0.8 to determine minimal sample sizes for all behavioral assays.