> Thus, population genetic, phylogenetic, and network analyses of nuclear sequences demonstrate that the giraffe is genetically well structured into four distinct species. This is consistent with divergence times of 1.25 to 2 million years ago (mya) among the four clusters (Figure 3C).
That genetic isolation is 100x longer than the human migration to the Americas. It's also 10x longer than the time to "Mitochondrial Eve" and about 8x longer than to "Y-chromosomal Adam".
Humans do not meet the phylogenetic criteria which justifies the use of "subspecies". See http://www.nature.com/ng/journal/v36/n11s/full/ng1455.html and many others. Quoting that reference:
> Modern human genetic variation does not structure into phylogenetic subspecies (geographical 'races'), nor do the taxa from the most common racial classifications of classical anthropology qualify as 'races' (Box 1). The social or ethnoancestral groups of the US and Latin America are not 'races', and it has not been demonstrated that any human breeding population is sufficiently divergent to be taxonomically recognized by the standards of modern molecular systematics. These observations are not to be taken as statements against doing research on demographic groups or populations. They only support a brief for linguistic precision and careful descriptions of groups under study. Terms and labels have qualitative implications.