MIGRATION — MEmEftTZttAGEN. 347 
ciently long days to collect a satisfactory supply of food for their 
offspring, or whether merely because the Arctic regions offer a more 
prolific food supply than more southern regions. If either of the 
two latter causes are correct, we should expect to find the Charad- 
riidae and Anatidae which breed in the Tropics to lay fewer eggs in 
the clutch than those which breed in northern Europe. We know 
that a plethora of food reflects itself on reproduction (c, f. snowy 
owls and rough-legged buzzards in lemming years in Scandinavia, 
and the increase of hyenas after wholesale deatlis among natives 
in East Africa). 
3. SPORADIC MIGRATION, INVASION, OR EXTENSIVE WANDERINGS. 
The well-known invasions of Syrrhaptes need no comment. That 
they would lead to eventual permanent colonization is almost cer- 
tain, but so far the species has never had a fair chance. There is 
no reason, however, why the wide distributions of Pterocles arenaHus 
or P. senegallus should not have been initiated by colonization after 
sudden invasion, for the sand grouse as a group are essentially wan- 
derers in search of suitable and rather specialized food, seeming to 
pride themselves on erratic movement and ignoring any seasonal 
lines of flight, which, generally speaking, constitutes migration; 
though in some spots they are particularly regular on migration, as 
is the case with P. arenanus of northern India. 
The rose-colored starling, aptly described as a veritable gipsy 
among birds, gives us a further illustration of colonization (in Italy 
and elsewhere) after invasion; and the various subspecific colonies 
of the crossbill {Loxia curvirostra) in the Mediterranean region 
might equally be due to colonization after eruption at some remote 
date, as opposed to either gradual expansion or regular migration, 
though the accuracy of such a theory to account for their present dis- 
tribution is by no means certain. 
4. HUMAN AGENCY, DIRECT OR INDIRECT. 
The introduction of such species as the pheasant, goldfinch, and 
starling to various parts of the world will suffice to illustrate ex- 
pansion of range due to direct human agency. In the case of the 
goldfinch, we find in the Bermudas that the bird has already estab- 
lished for itself a differentiation entitling it to subspecific rank. In 
the case of the introduction of the starling to Cape Town, it is inter- 
esting to note that the species has abandoned the migratory habit 
and has become a pure resident, not even congregating into flocks in 
winter. 
Contraction of distribution under this heading is the sad story of 
extermination, being generallj'^ confined to species having a very 
