92 BULLETIN 121, UNITED STATES NATIONAL MUSEUM. 



having come apparently from the direction of the South Pole; for, after in- 

 quiry of the captains of ships, I can discover no one who has met with them 

 between May and September, although their numbers at other times can be 

 computed only by millions. The obvious difficulty is that the season when tliey 

 disappear is midwinter. It is also noteworthy that the whole period from 

 September 17th till the beginning of May is taken up with the rearing of their 

 young. When they appear in September they are believed to have paired al- 

 ready, and they commence at once to scratch out their holes, the process last- 

 ing, with intervals, for six weeks. One bird is seen to be at work at a time, 

 and always in the night. In the daytime they depart seaward. Indeed, so 

 persistent is their desertion of the breeding-places by day that it would be 

 quite possible for unobservant persons to live on these islands and to be un- 

 aware that the petrels breed there at all, except perhaps at the season when 

 the young birds take to the water. It is almost true to say that I never saw a 

 petrel in the daytime near their rookeries during the eight years that I have 

 been visiting them. The universal belief is that these birds dare not alight on 

 the ground in daylight because of the gulls and crows, for the length of their 

 wings make them unable to fly off flat ground. The petrels depart about No- 

 vember 1st, and are only seen occasionally till November 20th, when they re- 

 turn in tens of thousands to lay their eggs, one for each pair, and they are laid 

 almost on the same night in each locality. Islands differ, some being earlier 

 than others, but I have heard of no egg being taken before November 18th. 

 If the egg is taken, there is good reason to believe that no other is laid ; but 

 the birds, being very gregarious, return to sleep on the rookery, although they 

 may have no young to tend. The Government permits consumption of the 

 eggs on the island, but none are to be exported. As a matter of fact, the eggs 

 are not fit to take after they have been laid three days. During the period of 

 incubation the parents take it in tui'ns to sit, exchanging positions after about 

 a week. The young birds appear about January 15th, and for a fortnight or 

 so tliey are in danger of snakes, which swarm on some islands, such as Chap- 

 pell Island and Babel, and are entirely absent from others, such as Little Dog 

 Island. The eggs, however, are safe, for the old birds can easily defend them- 

 selves when sitting. " Birders " say that birds and snakes are not found in the 

 same hole; but if a man seizes a snake in place of a bird he must pull it out, 

 for to relax his grip is to court danger. There are always signs if there is a 

 bird in a hole, for the parents carefully clean out the passage before they de- 

 part for the day and stuff the entrance with dry rookery grass. In March the 

 feathers of the young begin to grow ; before this they are fat, downy creatures. 

 By act of Parliament " birding " commences in Tasmanian lands on March 

 20th and continues till the surviving birds fly away about the beginning of May. 

 When the young petrels are fully feathered the parent birds desert them al- 

 together and depart seaward, the result being that the new generation is driven 

 from the holes by hanger, and, without assistance from the old ones, they have 

 to find their way to the water and to learn to fly and feed. All their traveling 

 is done by night, for fear of their enemies (the large gulls) ; but even so, a 

 great many are killed upon the water or upon the shore when too weak or in- 

 experienced to escape. Still, in spite of the efforts of man and of winged bi- 

 peds, the sea is black with young birds in May, and fully one-third of the young 

 petrels survive. 



Let us now imagine ourselves standing on a rookery in the evening awaiting 

 the arrival of the old birds as the sun approaches the horizon. The islands 

 where the petrels breed are as a rule devoid of trees. They are sandy knolls 

 covered with long grass, and seem useless, except to feed a few sheep. Chappell 



