Caste differentiation in social insects is a phenotypic plasticity for the high efficiency of individual tasks within a group. Morphological differences between the castes are caused by effects of neuro-endocrine systems and different gene expressions during larval stages. Recently, a lot of metabolic gene expressions between the castes have been determined and molecular mechanisms underlying caste differentiation of internal and external morphology have been gradually clarified. Sterile adult workers have a potential to change the reproductive states including reproductive organs and behaviors without any changes of external morphology under queenless conditions. Behavioral transition to reproductive individuals from sterile workers is caused by physiological changes in the brain and may eventually reconstruct the brain specialized to the caste specific behaviors. The basic physiological and molecular mechanisms of caste differentiation and caste transition seem common among species in eusocial Hymenoptera. Some of the hormonal functions for the caste transition, however, has been lost in highly eusocial species. In the present paper, I introduce common mechanisms underlying reproductive regulations among eusocial insects and discuss the evolution of endocrine functions for reproduction.

View full abstract

I describe some engineering methods that have been used frequently for estimating directions of arrivals (DOAs) of sound sources first. I further present a new DOA estimation method based on the generalized cross correlation function (GCC) weighted by the inverse of the difference spectrum between the channels. I evaluate the performances of the above methods assuming that two omni directional microphones are placed in free space. Next, I investigate the availability of the proposed GCC as a binaural sound localization model. The results of the investigation demonstrate that the performances of the proposed method in two-dimensional DOA estimation from binaural sound are comparable to those of the high resolution method Multiple Signal Classification (MUSIC) method which is considered to be a central method in engineering field. Moreover, the validity as binaural localization model is demonstrated by the performance in case where the phase information is omitted to simulate the auditory characteristics at the high frequency band.

View full abstract