128 THE RIDDLE OF MIGRATION 



their normal wintering grounds, until the gonads 

 have reached their minimum and then to liberate 

 them and expect them to stay, i.e. to prove as 

 sedentary as though they had completed their 

 migrations to the middle States. The hormone 

 producing tissue, of interstitial cells, is then want- 

 ing. If this tissue provides the stimulus to the 

 migratory impulse, the impulse should be entirely 

 absent at the minimal stage. The experiments 

 were run for four winters. During that time nearly 

 100 control juncos were liberated between Novem- 

 ber and February. Except for half a dozen that 

 were almost certainly killed by cats and northern 

 shrikes, every one was retaken in traps kept per- 

 manently set and baited outside the aviaries. In 

 spite of winter conditions, sometimes mild, at others 

 extreme, and the fact that the birds were far to the 

 north of their true wintering grounds, they showed 

 not the slightest interest in the possibilities of 

 escaping to southern latitudes. Some of them were 

 out and around for over two weeks before finally 

 being retaken in the traps. 



This is paralleled among wild birds. Thus mal- 

 lards that have stayed in various parts of the Prov- 

 ince till the end of November or December never go 

 south thereafter, no matter what weather may over- 

 take them. They will stay and starve to death 



