THK ROLLER 427 



of isolated trees in fairly open country, or resting on the telegraph 

 \\ires. It is not particularly shy, and can be approached within gun- 

 shot without difficulty. In the Gambia Kendall noticed that it would 

 dash down from the top branch* s of a tree after a lizard which had 

 rashly ventured into the open, but farther south it is generally 

 described as feeding on beetles, locusts, and grasshoppers picked up on 

 thf -i -omul. Von Ileuglin on our occasion came across many hundreds 

 in October feeding on the swarms of locusts. Being to a great extent 

 independent of water, and able to pick up a living in all but the most 

 arid deserts, as well as having considerable power of flight, the roller 

 is probably able to cross the Sahara without the necessity of diverging 

 in order to follow up the Nile valley or to keep near the coast Soon 

 after the beginning of April the first arrivals may be noticed in Tunisia, 

 and about the same time, or a little earlier, in Marocco and Algeria, 

 and here some settle down at once to breed, while their companions 

 pass over across the Straits of Gibraltar as well as to Sicily and Italy. 

 In Marocco it is a somewhat local bird, breeding among the ruins at 

 Larache and other places in considerable numbers. Mr. Meade- 

 Waldo describes how he encamped among some ruins where a colony 

 of quite three hundred of these magnificent birds were breeding. At 

 the time of his visit they were feeding their young with large and very 

 poisonous centipedes, quite six inches long not a particularly inviting 

 diet, to our ideas at any rate. The same writer also found it an 

 abundant species in the Great Atlas, and notes that it was breeding in 

 the old walnut trees at an elevation of GOOO feet In Algeria and 

 Tunisia it is also frequently found nesting in trees in wooded districts, 

 but the more usual breeding-place is in a hole in the steep bank of 

 some river or a crevice in a cliff face. Towards the end of April the 

 migration in Tunisia is at its height, and Mr. J. Whitaker 1 describes 

 how the telegraph wires lining the routes were constantly occupied by 

 small flocks of these birds, which kept on taking short flights and 

 perching again on the wires ahead. Curiously enough, in North out 



1 Bird* of Tunitin, ii. 52. 

 VOL. II. 3 I 



