NO. 3597 RODENT ETHOLOGY — EISENBERG 15 



sion and flexion of the body, or it may involve a movement of both 

 hind legs forward as the torso flexes. During the extension phase, 

 the chin may rest flat on the substrate with the result that the chest 

 i^ rubbed in the sand. The latter twisting is especially common in 

 the dipodids studied. 



The ventrum rub : In this study, the ventrum rub was defined 

 arbitrarily as extension and flexion of the body axis with the ventrum 

 pressed against the substrate. TypicaUy the hind feet are tlu'ust 

 forward even with or beyond either side of the head during the flexion 

 phase. The ventrum and perineal region are thoroughly scrubbed 

 into the substrate. It can be seen from this description that digging 

 and the side rub, plus the twisting of the body, will suffice to dust 

 the entire pelage with soil. The ventrum rub is not always necessary 

 as a pelage dressing component. 



Rolling on the back while writhing from side to side was rarely 

 shown by the species studied although it was observed occasionally 

 with Tatera indica, Meriones unguiculatus , and Pachyuromys chqjrassi. 



The frequency of side rubs to ventrum rubs varies in a species- 

 specific manner. Table 6 indicates the trends for seven species from 

 three families. The sandbathing Heteromyidae display a range of 

 ventrum rubbing from 23 to 45 percent, whereas the dipodid AUactaga 

 displayed only 8 percent ventrum rubs and the gerbils Meriones and 

 Gerbillus displayed virtuaUy no ventral rubbing as a component of 

 sandbathing. The discrete components of sandbathing occur in 

 sequences and each single event may be followed by a second event 

 within an interval range of one-half to an unknown number of seconds 

 exceeding 500. The frequency distributions for intervals separating 

 sandbathing acts were plotted. Since the majority of intervals fell 

 within 3K seconds, this interval was selected arbitrarily as the limiting 

 interval separating acts within the same sequence. For each species, 

 all sequential acts were plotted in a Latin square in a manner described 

 previously (Eisenberg, 1963c). Table 7 portrays the results, which 

 give an indication of how the components of sandbathmg are inte- 

 grated. Meriones and Gerbillus are prone to alternate side rubs from 

 one side to the other. AUactaga seems equally prone to alternate or 

 rub the same side a second time. Dipodom7/s deserti, Perognathus 

 calif orniciis , and P. 'parvus tend to alternate sides \\'hereas D. nitra- 

 toides shows a preference for rubbing the same side. All the Hetero- 

 myidae integrate ventrum rubs with side rubs into a functional 

 sequence. 



All species exhibit locus specificity in their sandbathing on con- 

 secutive days. There is also a pronounced tendency in some species 

 to sandbathe at the locus of a partner during a social encounter (see 

 p. 22). The occurrence of locus specificity and the exchange of loci 



