BREEDING OF GIRARDINUS PCECILOIDES. 907 



and I have thus been able to observe certain facts in relation 

 to their breeding habits, to one of which in particular I wish to 

 draw attention, namely that in some cases the male breeds before 

 the assumption of its secondary sexual characters. Before dealing 

 with this matter it is necessary to mention briefly the sexual differ- 

 ences, which are ver}'^ mai-ked. Females, which may measui'e up 

 to 25 mm. in length, are simply of a dull olive-grey in colour, 

 usually with a dark patch above the anal fin, while the males, 

 which do not attain quite so large a size, are most conspicuously 

 ornamented with red, blue, violet and yellow, with a dark ocellar 

 spot situated in the middle of the body, in front of or beneath 

 the dorsal fin, and another at the base of the caudal fin, these 

 spots appearing at least fourteen days before the brilliant hues. 

 Males are to be further distinguished from females by the 

 position and prolongation of their anal fin, which is transformed 

 into an intromittent pairing organ. 



The male is remarkably active, and is perpetually courting the 

 female, going through all kinds of antics in front of her, and as 

 breeding goes on all through the year, at least in captivity, the 

 latter is in an almost permanently pregnant condition, and within 

 a fortnight of having brought forth a brood (such a period 

 representing the duration of the gestation, at a temperature 

 of over 70""), once more brings a generation into the world. 

 It should be borne in mind, however, that a single impregnation 

 is sufficient for the fertilization of several broods, the embryos of 

 the second and third generations being already in an advanced 

 condition when the first young are born. Each brood comprises 

 from five to as many as twenty-five fish. 



The young fishes grow rapidly : about 4 mm. long at birth, they 

 double that length in a fortnight or so, when their sex can be 

 determined by the shape and position of the anal fin ; it is not, 

 however, until at least six weeks after birth that the males 

 become adorned with the brilliant colours which characterise the 

 adult, this period again varying according to the temperature of 

 the water. 



The first brood, containing a male which bred before the 

 assumption of its secondary sexual characters, was produced on 

 March 11th of this year, and numbered eleven. As is almost 

 invariably the case, a number of weaklings died within a few days 

 of their birth, and eight fish remained at the end of a fortnight, 

 when I found the sexes to be equally divided, which is rem.ark- 

 able, for as a rule females outnumber males by about three to one. 

 On May 7th, six young ones were born of this brood, and at the 

 time none of the males had developed their adult coloration, the 

 only markings distinguishing them from the females being the 

 spots beneath the dorsal fin and at the base of the caudal, and it 

 was in fact not until six days later that one of these acquired its 

 full sexual livery. The others followed suit on May 17th and 

 18th. These broods were kept at an average temperature of 80°. 



In another case, a bi-ood of fourteen fishes, born on June 5th, 



