Shrinking erectus

Ann Gibbons reports on the AAPA meetings with a story about all the Homo erectus pelvis and stature papers (“Human ancestor caught in the midst of a makeover,” subscription required). Research on the proportions of early Homo was the main event of the meetings, and Gibbons really caught the highlights of the story.

I wrote about body size in Homo erectus a few months ago, and much of the story follows from the basics I outlined there (“The changing height of Homo erectus). But there I emphasized that the estimated adult height of KNM-WT 15000 was an outlier in a relatively small body size distribution.

What I didn’t anticipate is that some interesting work might come along to question the tall adult stature estimate for that skeleton. Gibbons describes the work of Ronda Graves and colleagues, presented at the meetings:

Using intermediate growth rates, graduate student Ronda Graves of Stony Brook University in New York state calculated that Nariokotome Boy would have had less time than originally predicted to reach his adult height when he died. She estimated at the meeting that he would have reached 163 cm in height and 56 kg in weight as an adult"shorter and wider" than previously thought.

This seems very short, at least when I first saw it. On reflection, Ohman and colleagues (2002) had provided a stature estimate at death of KNM-WT 15000, as only 147 cm, and they suggested it might have been as short as 141 cm. That’s an awful lot shorter than had previously been estimated on the basis of regressions.

If Graves and colleagues are right about the lack of a human-like growth spurt, an additional 20 cm (8 inches) wouldn’t be unusually small for an adult stature. Those stature estimates would put KNM-WT 15000 between the 50th and 90th percentiles for American 10-year-old boys, or between the 25th and 75th percentiles for 11-year-olds. By contrast, an adult stature of 163 would be around the 3rd percentile for adult American men. The assumptions about growth totally determine the outcome for adult height.

The credibility of the growth assumptions can only be tested by looking at other adult and juvenile remains. There is much more to say on this topic, but I’ll point out one relevant comparison: The estimated stature of the adult skeleton from Dmanisi, including the complete D4167 femur and D3901 tibia, is between 145 and 166 cm. Graves’ KNM-WT 15000 stature estimate is right within this range.

Meanwhile, there was a lot of disagreement about hips.

[Scott] Simpson and Linda Spurlock of the Cleveland Museum of Natural History realigned the pieces of Nariokotome Boy's pelvis, guided by a female H. erectus pelvis from Gona, Ethiopia, that Simpson reported 2 years ago (Science, 14 November 2008, p. 1089). They found that the widest measure from side to side on the boy's pelvis is 255 to 260 millimeters rather than 225 to 230 mm. This would give the boy an adult hip breadth of 295 to 301 mm rather than the 266 mm originally proposed, and would match those of the short, wide-hipped female from Gona, whose pelvic breadth was 288 mm. "H. erectus was not simply a small-brained modern human," says Simpson.

Simpson’s reconstruction seemed reasonable, and it’s actually not that big a difference – roughly an inch and a half (3 cm) in bi-iliac breadth. The main differences were in the overall shape of the pelvis, being shorter with a more flaring iliac blade.

Gibbons describes the disputation that happened after Chris Ruff’s presentation. Ruff has suggested that the Gona pelvis may not represent Homo – that its broad proportions and small acetabula (hip sockets) suggest it may have belonged to an australopithecine (presumably, A. boisei).

Much of the disagreement comes down to the estimation of femur head diameter from acetabulum breadth – Ruff (2010) gave an estimate of 32.6 mm, Simpson and colleagues estimated between 35 and 36 mm, based on a different method. What you would want is enough acetabula of both genera to be able to examine their variation directly. We don’t have such a sample; what we have are a few acetabula and several femur heads. We have the additional problem that living people seem to have a different relation of femur head and acetabulum diameters than in other anthropoids, and it’s not obvious which should be applied to early hominins.

I guess (in the relative absence of data) that this acetabulum diameter of the Gona pelvis was in the zone of overlap between Homo and Australopithecus. There’s no question that later Homo – say after 1 million years ago – is substantially larger in acetabulum diameter, from every specimen so far described. But there are occasional small specimens of Homo even in the Middle Pleistocene. At 1.15 million years old, the Gona specimen is more than 300,000 years later than the last known occurrence of Australopithecus. The femur head that would fit the Gona acetabulum would be smaller than KNM-ER 1472 or D4167 from Dmanisi, both around 40 mm. At least one australopithecine femur head (AL 333-3) is that large, so the femur head diameter distributions do overlap. The STW 431 acetabulum diameter is a sliver larger than that of the Gona pelvis (Ruff 2010 makes it 3 mm bigger, but other workers have given a smaller estimate). SK 3155 may well be Homo and has a smaller acetabulum.

Of course, if we go as far as SK 3155, we have to consider the topic of the Malapa innominate. Can we tell small-bodied Homo from Australopithecus on the basis of pelvic morphology? Several people writing about the Gona pelvis have made it sound like a bigger version of Lucy’s. But that’s not really true. The australopithecine-like appearance comes from its breadth and consequent features, including the long pubes and flaring anterior ilia. The rest? Maybe there’s something here for a clever anatomist.

UPDATE (2010-04-27): I have some e-mail about the last occurrence of A. boisei, which I wrote above was more than 300,000 years older than the Gona pelvis.

The most potent counterargument is Swartkrans Member 1, which has uranium-lead dates around 830,000 years ago, and has been placed by many workers around a million years ago. I actually hadn’t been thinking of South Africa. But it is relevant, as the East African record between 1.4 and a million years ago may not be strong enough to argue that the last occurrence of A. boisei is really very close to the extinction time.

Meanwhile, there is OH 36, an ulna from Olduvai Gorge that may represent A. boisei. Since it’s (obviously) not cranial, and is quite large and robust compared to postcranial remains that are associated with A. boisei, I’ve always been very skeptical of that assessment. If there’s one feature of the ulna that actually has some phylogenetic importance in the Early Pleistocene, I figure it’s size.

But given the current question about body size, that reason for skepticism may have receded in importance. On the other hand, OH 36 seems to represent a substantially bigger individual than the Gona pelvis, so maybe introducing robust australopithecines into the mix doesn’t help anything.

Several things puzzle me. Even into Member 1 times, Swartkrans is dominated by A. robustus, with very little Homo. In East Africa, A. boisei is never quite so predominant in the hominin assemblage as the case in South Africa, but was nevertheless very common up to 1.5 million years ago. Did it persist much later? Was it cryptic from the point of view of the fossil record? Are the Swartkrans dates older than we think?


Gibbons A. 2010. Human ancestor caught in the midst of a makeover. Science 328:413. doi:10.1126/science.328.5977.413

Ohman JC, Wood C, Wood B, Crompton RH, Günther MM, Yu L, Savage R, Wang W. 2002. Stature-at-death of KNM-WT 15000. Hum Evol 17:129-141. doi:10.1007/BF02436366

Ruff C. 2010. Body size and body shape in early hominins -- implications of the Gona pelvis. J Hum Evol (in press) doi:10.1016/j.jhevol.20 09.10.0 03