The Flores Hominid and the Evolution of the Shoulder

Homo floresiensis more widely known as the "Hobbit," may have had arms that were very different from those of modern humans.

A paper in the current issue of the Journal of Human Evolution explores the anatomy of H. floresiensis. To explore this we first have to understand the concept of "Humeral torsion." Humeral torsion is the orientation of the humeral head relative to the mediolateral axis of the distal articular surface. Don't bother reading that sentence again, I'll explain it.

The humerus is the upper arm bone, that runs between your shoulder and your elbow. The humeral head is the round ball-like part that fits into the "shoulder" ... if you have a dislocated shoulder this is the part that is not where it is supposed to be. It articulates with the clavicle and the scapula, but really, all three of these bones are held together with a mass of connective tissue. The mediolateral axis of the distal humerus is the plane defined by your wrist. If you balance a pencil on your wrist, the pencil is more or less resting in the plane of your humeral (arm bone) distal (far end of the arm bone) mediolateral axis.

So, if your arm is at rest, the humeral head is more or less pointing towards the middle of your body. If, when your arm in a nice relaxed state your wrist is parallel to the side of your body, then the difference between the orientation of the humeral head and the distal humeral bit is minimal. If, however, when you stand at wrest your hand tends to point backwards, like if your arms were extra long you'd be dragging your knuckles, then the difference between these two orientations is large, like maybe 90 degrees or so. (Some anatomists call this angle, the 90 degree torsion, zero degrees and measure deviance from that number. Others call it 90 degrees and measure from that number. It is all a matter of style.)

Most mammals, and this is the presumed primitive condition in mammals, have a humeral head that faces towards the back of the body, with the humerus sticking down towards the ground (mammals, generally, are quadrupedal so the humerus is a leg bone, not an arm bone!). Apes, however, have evolved a humerus with its head pointing more toward the middle of the body. Humans are pretty much the same as all the other apes at a gross level. However, within the apes, modern humans are "more different" than the other apes, with a fairly high degree of torsion.

To put it simply but clearly, the ape humerus is twisted a certain amount, an the modern human humerus is twisted a slightly different amount. The differences are small, but fairly consistent. The following graph shows the pattern of humeral torsion across a number of human samples, Neanderthal, and some early hominids. Flores is LB1/50. There is clearly a lot of overlap, but by grouping all of the early hominids we see that there may be a pattern, and that Flores may fit into the early hominid pattern better than the modern human pattern

Fig. 4. Mean humeral torsion angles plus 95% confidence intervals for samples (CIs) from literature sources for different modern human and fossil hominin groups, and torsion measurements for individual fossils.

This vartiation also plays out in other aspects of the bones. The techniques used by the researchers to investigate this involves sclaing to body size by plotting the length of the humerus and the length of the clavicle, as shown here:

Fig. 6. Scatter plot of mean clavicular length against mean humeral length in nonhuman primates, a variety of modern human groups, samples of early modern Homo and Neanderthals, and LB1 (Flores) and KNM-WT15000 (a Homo erectus).

As the humerus gets longer, so does the clavicle. This relationship across primates (including prosimians and New World monkeys, as well as Old World Monkeys and apes) follows a pattern from which there are a few divergences. Ateles and baboons have short clavicles, while some of the apes and all humans have long clavicles. One problem with these data is that the range of variation increases as one gets to the larger bodied animals. Is this because the variation itself is scaled to body size? That is fairly likely. But it is also true that the nature of the locomotion and positional behavior is diverse in the upper range, with gorillas being ground-doweling quadrupeds who knuckle walk and orangs having a virtually unique arboreal form of locomotion owing to their body size. However, it is apparent that the modern humans and Neanderthals are way "off the line." Look close, you'll find LB1, the Flores sample, which is dead on the line for primates in general. If all I knew about Flores was that it was small, had an ape-size brain, was bipedal and lived in southeast Asia, then I saw this graph, I'd ask "So, you think Flores is a ground dwelling gibbon? Interesting, there are no extant forms of this kind of gibbon, you've got something very interesting there..."

This study looks at a number of other factors, but the same conclusion is arrived at again and again. The Flores humerus is more like a generalized ape humerus at the broad scale, but it is most like an early hominid, in particular, a Homo erectus humerus in details. Again and again, Flores keeps turning out to be a miniaturized Homo erectus, but with enough differences to ask if the similarities are convergences or shared shared ancestral traits.

The authors suggest:

The unexpected combination of primitive and derived characteristics of H. floresiensis and early H. erectus shoulder material highlights our ignorance regarding the course of transformation of the hominin pectoral girdle and shoulder from a more ape-like ancestral condition to the morphology of modern humans.

Several hypotheses are suggested to account for the Flores/erectus pattern, and the subsequent shift to the modern human form in one lineage (not Flores, presumably). The reorientation of the scapula and its relationship to the humerus could be an adaptation to facilitate stone tool manufacture. The reduction of overhead hanging and grasping in connection with arboreal locomotion could also be indicated in early hominids, and the further reorientation in modern humans could be related to throwing. In other words, Flores would not have been throwing spears or rocks, or even baseballs, like modern humans would. Personally, I thik the initial shifts we see could relate to using spear-shaped implements to dig, with the later shifts involving using similar implements to both dig and hunt. As usual, running is also thrown into the mix: The modern human shoulder is better adapted to efficient running than the Flores or erectus shoulder.

The authors conclude:

Debate continues regarding the proper interpretation of the Flores hominins ... Although the controversy may continue until additional material, especially new skulls, are found, studies looking beyond brain size ... have observed unexpected morphology that defies simple explanations. Whatever the ultimate taxonomic attribution of the Liang Bua hominins, their unique morphology suggests unforeseen diversity in the human family. In regard to the present study, while LB1 and the Nariokotome [Homo erectus] skeleton differ in many ways and are known from very different times and places, they are similarly distinct in displaying a relatively short clavicle and low humeral torsion. We believe these are not chance similarities, but part of a previously unrecognized functional complex that characterized early H. erectus and was retained in H. floresiensis.

And I'll let them have the last word except on one point: Increasingly, it becomes possible to consider time and space in sussing out these evolutionary patterns (two things that are found very far from each other are less likely to be closely related than two things found near each other ... in time or space). Fine. But using this logic too freely may result in misunderstandings. On one hand, one takes the risk of, essentially, asking the question: "If humans evolved from a chimp, why are there still chimps.." and o the other hand, one ignores the potentially dramatic events that can happen in biogeography. The zebra is more closely related to some Asian horse than it is to the springbok it is standing next to in the savanna.

Larson2007.    Larson, S.G., Jungers, W.L., Morwood, M.J., Sutikna, T., Jatmiko, Saptomo, E.W. et al.. Homo floresiensis and the evolution of the hominin shoulder. J Hum Evol 53, 718-31(2007).