One of the phenomena investigated by P. W. Geier and co- workers at the CSIRO Division of Entomology, Canberra, in a series of studies (Geier and Springett (1976), Geier and Briese (1977), Geier and Oswald (1977)) of the light-brown apple moth, Epiphyas postvittana (Walker), was the tendency for males to be consistently less abundant than females in samples drawn from Australian field populations. The existence of a heritable condition (‘Q-condition’) in ‘carrier’ females of the light- brown apple moth, which caused them to produce predominantly female progenies, was established by Geier and Briese (1977). A further extensive investigation of the Q-condition was carried out by Geier and Briese in 1978. In this connection they consulted me (at the time a visiting member of the CSIRO Division of Mathematics and Statistics at Canberra) with regard to the modelling of some aspects of their data, and I give an account here of the statistical work (published as part of Geier, Briese and Lewis (1978)) which I carried out in support of their research. Apart from its statistical content, it perhaps has some interest as a cautionary tale about the pitfalls which a statistician may encounter in offering too dogmatic an interpretation of a researcher's results.

The problem

The problem was outlined as follows in a letter to me from Dr D. T. Briese, quoted here with his permission: