REED WARBLER 



to speak, one accustomed to construct their nests on old and 

 the other on new reeds. And, if this be so, it will to some 

 extent explain the peculiarly erratic arrival of different 

 individuals even in the same reed bed, the earlier comers 

 being those which had adapted themselves entirely to the 

 use of dead reeds. The actual number of stems to which 

 the nest is woven varies very considerably; one bird will 

 make use of three reeds, while another of no less than seven. 

 But the number chosen must have been in the past, and 

 must be still, so regulated as to ensure the safety of the 

 young ; and in a case of this description we need have little 

 hesitation in attributing the result to selection. A nest woven 

 to two reeds, if it were possible to construct it thus, would not 

 be durable, and even when woven to three does not always 

 appear to be sufficiently secure, yet this number is sometimes 

 resorted to ; and if three reeds be really sufficient, why, in 

 so many cases, should four, five, six, or even seven be used ? 

 It may be suggested that where a small number of stems is 

 used it is the first attempt of a young bird, and that as it 

 finds the number insufficient to hold the nest securely, it 

 profits by experience, and at the next attempt adds another 

 and yet another stem. But to profit by experience, persistence, 

 with varying effort, is necessary. " Eepetition of trials with 

 variation of procedure is a sort of perceptual experiment. 

 The results of previous experiment determine and facilitate 

 future action, inasmuch as unsuccessful modes of procedure 

 are gradually eliminated and successful modes alone survive." 1 

 A bird builds a nest but once a year, and therefore the con- 

 ditions necessary are not really present in a case of this 

 description, and to expect it to realise the cause of the failure 

 of its first attempt, and to overcome its difficulty in advance, 

 would be to raise the level of its whole behaviour from the 

 perceptual to the ideational plane, which I do not for one 

 moment believe we should be justified in doing. And whereas 



1 Stout's " Manual of Psychology." 

 43 



