80 NATURAL HISTORY 



than elsewhere : is it owing to the vast massy buildings of 

 that place, to the many waters round it, or to what else ? 



When I used to rise in a morning last autumn, and see 

 the swallows and martins clustering on the chimneys and 

 thatch of the neighbouring cottages, I could not help being 

 touched with a secret delight, mixed with some degree of 

 mortification ; with delight to observe with how much ardour 

 and punctuality those poor little birds obeyed the strong 

 impulse towards migration, or hiding, imprinted on their 

 minds by their great Creator ; and with some degree of 

 mortification, when I reflect that, after all our pains and in- 

 quiries, we are yet not quite certain to what regions they 

 do migrate ; l and are still farther embarrassed to find that 

 some do not actually migrate at all. 



These reflections made so strong an impression on my 

 imagination, that they became productive of a composition 

 that may perhaps amuse you for a quarter of an hour when 

 next I have the honour of writing to you. 



LETTER XXIY. 



TO THOMAS PENNANT, ESQUIRE. 



SELBORNE, May 29, 1769. 



HE Scarabceus fullo 2 I know very well, 

 having seen it in collections ; but have 

 never been able to discover one wild in its 

 natural state. Mr. Banks told me he 

 thought it might be found on the seacoast/ 

 On the 13th of April, I went to the sheep-down, where 



1 The winter haunts of the swallows are now well known to naturalists. 

 See an article on swallows, in continuation of a series on " Our Summer 

 Migrants," published by the writer in the "Field" of October 14th 

 and October 21st, 1871, in which a detailed reply is given to the 

 inquiries, "What do swallows feed upon?" and "Where do the 

 swallows go in winter?" ED. 



2 Melolontha fullo, Fabr. an insect of the same genus as the better 

 known cock-chafer Melolontha vulgaris. 



3 Mr. Bennett has remarked that all the specimens of this fiiie chafer 

 that have yet been captured in England, and they are very far from numer- 



