No. 488] 



HABITS OF THE SHREW 



507 



with three openmgs. An extension of 40 cm. had Keen made from 

 one of the former openings. 'VhU ^andy M>il liad not tu'en packed. 

 Hence the difficuhy in hnrrowinu in it mn>t liavc l,cen dne to its 

 weight, not to its hai-dness. In neither the l.h.ck nor >andy soil 

 did the shrew loosen the soil with its teetli as k'eimicott ( I SoT. j). 

 94) has conjectured Neither then nor at any other time (hn-ing 

 the confinement of tlie shrew did I observe any marked surface 



Food. 



