136 DENDROCYGNA BICOLOR 



ruary 9. Hartert and Venturi (1909) speak of a ripe egg having been taken from a 

 female on November 3, in northern Argentina. Immature birds were seen on the 

 Shire River in Portuguese East Africa in October (Hartert, 1898). In India they 

 breed inmost places in July and August; in Nadia, at the end of June, and in Rung- 

 pur principally in August, with a few in September. 



There are very few minute observations as to display and the details of the nesting 

 operations. Pinioned specimens are so inactive, and are so seldom disposed to breed, 

 that there are almost no notes from the zoological gardens. Finn (1919) observed no 

 marked display before the pairing act, but noted a postlude in which both parties 

 executed a step-dance in the water, with one wing held aloft. It is true of the Tree 

 Ducks as a whole that, so far as is known, they have no marked display beyond the 

 reciprocal dipping and lifting of the head and neck, mentioned by Heinroth (1911). 



The nest is by no means invariably placed in trees. In California it is almost ex- 

 clusively on the ground, and so far as the evidence goes, the same is true of Mexico 

 and the treeless regions of the Argentine. In the tropical portions of South America, 

 in India, and probably in Africa the birds make use of trees, depositing their eggs in 

 the abandoned nests of other birds, or occasionally in hollows. Baker (1908) found 

 nests in India which he believes were constructed by the birds themselves; but I am 

 by no means convinced that any of the ducks are capable of building large and elab- 

 orate nests in the branches of trees. Very instructive is Finn's (1915) remark that 

 in India they occupy old nests, holes in trees, or suitable boughs, on which they 

 make a nest of their own, and that as yet they have never been found nesting on the 

 ground. The trees selected are usually small ones located on tiny islands in jheels, 

 or overhanging the water. Baker (1908) says the nests are very roughly constructed 

 of twigs, sticks, and grass, and in a few cases covered with a dirty mass of weeds. 

 They average some eighteen inches across and are placed, not so often in forks, as 

 on tangles of branches, or where the first big branches spread from the bole of a 

 large tree. One nest was in the crown of a date palm, while in another section of the 

 country, Rungpur, they were found selecting large trees, and building thirty feet 

 or so from the ground. Careful observation to see how much actual nest construction 

 is carried out would be extremely interesting in these cases. 



Where they nest on the ground they choose either marshy localities or dry land. 

 In the former case, in the Argentine, the nests were built in dense flag-beds, and con- 

 structed of dry flags. When in the grass, they were unlined and it is indeed doubtful 

 whether the nests are ever really lined like those of true ducks (Gibson, 1920). 

 Other authorities found nests made of stems and leaves, on the water among reeds 

 (P. L. Sclater and Hudson, 1889). The best account of the nesting habits of these 

 ducks in California is that of Shields (1899), who discovered many nests in a large 

 tule swamp, where they were formed by breaking off the adjacent stalks. These nests 

 were usually, but not always, well hidden, but as compared to the preceding species 

 the Fulvous Ducks are evidently less careful about concealing the nests. 



