Summary Psychosocial factors are thought to influence risk and survival from cancer. We have previously studied specific behaviours in transgenic male CD-1 MT42 mice, which overexpress the gene encoding human transforming growth factor a (TGFa) in multiple tissues, and which Transforming growth factor a (TGFa) is a polypeptide which exhibits approximately 35% sequence homology to epidermal growth factor (EGF), and interacts with cells through the EGF receptor. Both TGFx and EGF can act as potent mitogens in a number of epithelial cell systems (Carpenter & Cohen, 1979). It has been postulated that TGFa plays an important role in neoplastic transformation, since it is highly expressed along with its receptor in many TGFa growth responsive tumour cells (Bates et al., 1988;Derynck et al., 1987;Rosenthal et al., 1986;Watanabe et al., 1987). The results of studies with transgenic CD-1 mice, in which chronic overexpression of TGFa is directed by the metallothionein (MT) promotor to multiple tissues throughout their life (Jhappan et al., 1990;Sandgren et al., 1990), have shown that male mice bearing a TGFa transgene develop a high incidence of spontaneous hepatocarcinomas at the age of 10-15 months (Jhappan et al., 1990). Other abnormalities in these animals (stomach, pancreas, etc.) have emerged as studies have proceeded.Human and animal data suggest that an ability to cope with stress influences vulnerability to develop cancer (Hilakivi-Clarke et al., 1993b;Holland, 1989;Sklar & Anisman, 1981). Perturbations in natural killer (NK) cell activity (Chuang et al., 1990;Saibara et al., 1989;Shirai et al., 1990) and steroid hormone levels (d' Arville & Johnson, 1990;Yager & Shi, 1991) may mediate the effects of psychosocial factors in cancer. We have utilised transgenic male MT42 TGFax mice in studies assessing those behaviours associated with tumourigenesis (Hilakivi-Clarke et al., 1992a). When compared with age matched non-tran sgenic control CD-1 mice, 2-3-month-old male transgenic TGFa mice spent significantly longer times immobile in Porsolt's swim test, and in aggressive behaviour (Hilakivi-Clarke et al., 1992a). The male TGFa mice also exhibited a 25% lower NK cell activity, and a 4-fold increase in the plasma levels of 17p-estradiol (E2) than the controls (Hilakivi-Clarke et al., 1992 al., 1990). This is surprising, since TGFa appears to play a significant role in mammary tumourigenesis (Bates et al., 1988;Liu et al., 1987;Sandgren et al., 1990), and female mice in other transgenic TGFa models exhibit an increased incidence of mammary tumours (Matsui et al., 1990;Sandgren et al., 1990;Stuart, 1984). The expression of TGFoa mRNA appears equally elevated in both sexes of MT42 mice (Hilakivi-Clarke et al., 1993a;Jhappan et al., 1990). To investigate behavioural changes associated with the overexpression of TGFa in the female mice, we utilised (i) Porsolt's swim test, which is thought to measure both depressive behaviour and an animal's ability to cope with stress (Garcia-Marquez & Armario, 1987;Hilakivi et al., 19...