Variation in blood flow mediated by the posterior communicating collateral arteries (PComs) contributes to variation in severity of tissue injury in obstructive disease. The “extent” of PCom collaterals (lumen diameter &/or number—ie, presence bilaterally, unilaterally or absent) varies widely in humans and other species due in part to developmental differences, since well-known patterns of variation are present at birth. However, mechanisms underlying this variation have not been examined. Herein we used angiography to probe involvement of genetic, environmental and stochastic factors. PCom extent varied widely among 7 mouse strains with different genetic backgrounds that was unrelated to differences in PCA diameter or brain weight. The pattern differed strongly from that previously reported by us for differences in extent of pial collateral arterioles. Like pial collaterals, PCom diameter was reduced by advanced aging (-28±3%, n=11-16, p<0.01; opposite seen for PCA diameter) and hypertension (-14±8%, n=7-9, p<0.05; -18% for PCA diameter). In contrast to pial collaterals, obesity, hyperlipidemia, metabolic syndrome and diabetes had no effect on PCom extent (or PCA diameter). Intrauterine growth restriction reduced PCom diameter 55±4% (n=6-17, p<0.05; -24% for PCA) in the adult despite catch-up growth of brain after birth, while pial collaterals were unaffected. Compared to pial collaterals (and PCA), PCom extent evidenced >8-fold larger stochastic variation (n=16, p<0.01). In addition, PComs underwent outward remodeling after MCA occlusion that varied with genetic background (12-133% lumen enlargement among 5 strains; n=5-14, p<0.05-0.001) and was greater on the side of occlusion in most strains. In conclusion, variation in number and diameter of PCom collateral arteries arises from genetic, environmental and stochastic factors. The responsible genetic polymorphisms appear to differ from those that contribute to variation in extent of pial collateral arterioles. Extent of PComs is also adversely affected by intrauterine growth restriction, aging and hypertension. These findings suggest possible sources of PCom variation in humans and provide information relevant when studying mouse models of occlusive cerebrovascular disease.