Perhaps the best-known fossil of Homo erectus is the one known as the Nariokotome boy (or Turkana boy) - a boy who, when he died at around 9 years old, already stood nearly 160cm tall. Members of this tall, long-legged species are generally regarded as being the first of our genus to move out of Africa and into Europe and Asia. Fossil remains from Dmanisi, in Georgia, suggested that this migration must have occurred at least 1.8 million years ago. However, the remains consisted only of skulls, and so nothing was known of what the Dmanisi individuals might have looked like - until the recent publication of a description of the post-cranial remains.
The new Dmanisi descriptions are summarised by Gibbons (2007), from a full descriptive paper in the journal Nature. The original paper describes the post-cranial bones from both teenage and adult individuals. All were shorter than the Nariokotome boy - at the same age they would perhaps have stood shoulder-high to him. Altogether, the Dmanisi remains come from people who were between 145 and 166cm tall, and their crania are much smaller than those from other erectus fossils. In addition, their shoulder & upper arm bones resemble those of australopithecines (and of H. floresiensis). While some scientists regard the Dmanisi fossils as representatives of H. habilis, or the new species H. georgicus, the researchers who described the remains feel that they represent early erectus, based on the proportions of their bodies and legs.
What can we take from this? One thing is that fossils (like modern humans!) are quite variable. Apart from anything else, this means that there will always be debate about the names and relationships of many fossil specimens. (After all, you can't apply the standard biological species defintion to a fossil - there's no chance of them breeding!) Lieberman (2007) comments that even by modern standards there is unusually high variability within fossils attributed to Homo erectus: apart from the differences in height, cranial capacities in early specimens vary from around 600cc in the Dmanisi individuals, to nearly 1100cc in 1-million-year-old specimens from Africa, and even more in younger fossils. Some of the variation may be due to sexual dimorphism, some to evolutionary change in a species that spans a million years of human prehistory. Combined with recent evidence of a substantial temporal overlap between erectus and habilis, this suggests there's a lot we still don't know about our evolutionary family tree.
A. Gibbons (2007) A new body of evidence fleshes out Homo erectus. Science 317: 1664
D. Lieberman (2007) Homing in on early Homo. Nature 449: 291-292