2018-06-25
The giraffe (Giraffa camelopardalis) is a marvelous animal, better known than its forest-dwelling relative the okapi and countless other charismatic artiodactyls in large part due to its proportions. As a result, there has been no shortage of speculation about the origins of the giraffe's iconic neck.
Making things up about giraffes is a well-established tradition: Timotheus of Gaza (~500 CE) reports that "the giraffe is an Indian animal; and it is born from the intercourse of different animals" (presumably a camel and a leopard, given the name). Lamarck also had his hypothesis, but there is enough confusion around Lamarck already that I'm not really excited about engaging with that without proper caution (to paraphrase a paleontologist friend of mine: "Almost no one knows what Lamarck said, Lamarck said crazy s**t").
There are two proposals in particular that have stood out in the past few decades: the more traditional feeding hypothesis first proposed by Darwin (Padian and Horner, 2010) and the sexual selection hypothesis proposed by Simmons and Scheepers (1996). This is a fun topic, but it also provides an interesting case study on the role of trends and novelty in the popularization of ideas.
Darwin's hypothesis is straightforward and familiar. He reasoned that longer necks conferred survival advantages by enabling longer necked individuals to forage in less competitive domains, leading to greater survival rates and fitness in longer necked individuals. The Simmons and Scheepers hypothesis was proposed in the context of the resurgence of interest in the role of sexual selection that occurred in the 1990s. They correctly noted that empirical evidence for the feeding hypothesis was inadequate, noted the importance of necks in male giraffe sexual competition, and appealed to observed sex differences in neck morphology. The idea was novel and exciting, and quickly entered textbooks (Freeman and Herron, 2004).
We should clarify exactly what is meant by sexual selection. Padian and Horner (2011, p. 1), based on Darwin’s definition, define sexually selected features as those “features ... present in one sex and not the other, and that ... either attracted mates or repelled rivals”, specifically adding that for a particular feature to be sexually selected it needs to have “not simply allometric sexual differences”. By this standard, in order for it to be likely that the giraffe’s neck length was sexually selected, there would need to be a demonstrated difference with positive allometric scaling of neck length in males being greater than that of females. Simmons and Scheepers (1996, p. 777) cite work done by others finding that female giraffe necks are on average between “30-40 cm” shorter than males, but this says little about neck differences in particular as female giraffes tend to be substantially smaller than males in general (Simmons and Scheepers, 1996, p. 782). Rather than measuring neck length directly, Simmons and Scheepers (1996) instead used neck mass as a proxy measure assuming that it would correlate with neck length and found that there was positive allometric scaling with neck mass in males but isometric scaling with neck mass in females as body weight increased.
This seems, prima facie, to be reasonable evidence for a role of sexual selection in the evolution of the giraffe’s elongated neck, but there are complications. The claim is that neck length in particular was sexually selected, so the proxy measurement of neck mass may not reliably indicate sexual dimorphism in the feature of interest. Indeed, when the analysis was carried out with direct neck length measurements, Mitchell et al. (2009, p. 278) found that “throughout growth, the necks of female giraffes elongated faster than their legs and faster than was the case in male giraffes” and that “at their mature body mass of 1100 kg, the neck length to foreleg length ratio in our females was 1.13 whereas in 1500 kg males it was 1.08”, which shows the exact opposite pattern predicted by Simmons and Scheepers (1996) with females having proportionately longer necks. Mitchell et al. (2009, p. 282) suggest, in reference to the difference in conclusions to Simmons and Scheepers (1996), that “the main reason why our conclusions differ is that we measured neck length and leg length whereas in their study these were not measured”.
There are a few other curious differences between the results of Simmons and Scheepers (1996) and Mitchell et al. (2009) that are worth mentioning. Simmons and Scheepers (1996) found that male head mass and neck mass continued to increase linearly throughout an animal’s life, whereas female neck mass tended to level off at a certain age. In contrast, Mitchell et al. (2009) found no such difference between the sexes. Unlike the different patterns seen in neck length and neck mass, this is a bit harder to explain. Mitchell et al. (2009, p. 278) mention that such discrepancies may result from different growth patterns in Namibian and Zimbabwean giraffes but find the possibility “unlikely”. Mitchell et al. (2009) also suggest that differing size and maturity ranges in the samples studied may have contributed. Neither of these explanations is truly satisfactory and further work would need to be done to settle the matter, ideally on several regional varieties of giraffe and drawing from the full size distribution of the animals. Still, as the matter stands, the anatomical evidence seems insufficient to support sexual selection as a large factor in the elongation of the giraffe’s neck, but discrepancies in observations between Simmons and Scheepers (1996) and Mitchell et al. (2009) suggests that further anatomical studies are required before it can be ruled out.
Simmons and Scheepers (1996) also reference behavioral observations of male giraffes done by Pratt and Anderson (1985) and conclude that neck size is more important than overall size in determining dominance and mate access. Simmons and Scheepers (1996, p. 779) write “in some cases ... [dominant] bulls were smaller than [non-dominant] bulls but were ranked as [dominant] because of larger necks”, concluding that this supports the notion of sexual selection on neck length. However, once again, they seem to be conflating neck size and neck length without proper justification. Indeed, Pratt and Anderson (1985, p. 771) specifically state that the “important [difference is] the stouter neck”, saying nothing about length despite earlier in the paper referencing the importance of ossicone length in determining dominance. Pratt and Anderson (1985, p. 771) did, however, find that in general dominant bulls “tend to be larger”, very much in line with the general pattern observed in ungulates. Simmons and Scheepers (1996), overall, presented the behavioral evidence in support of the sexual selection on neck length as being stronger than it actually is.
Let's now consider the traditional explanation, often called the feeding hypothesis, which may also not be the entire story. Simmons and Scheepers (1996, p. 775-776) reviewed the literature on giraffe feeding behavior and found that giraffe browsing usually occurs at or near shoulder level with much of the browsing happening at heights accessible to other animals, suggesting that such behavior points to an absence in selection that could maintain or produce neck elongation. Taylor et al. (2009, p. 5), in response to this line of thinking, write that “even if it were true that giraffes habitually browsed with their necks horizontal whenever possible, it would remain the case that the long necks, enabling higher browsing when necessary, would provide access to scarce food during times of environmental stress”. This seems to be a plausible explanation, but it is not supported by the evidence. Mitchell et al. (2010) analyzed the demographic profile of giraffe deaths in response to drought conditions and found a distinctly increased chance of death among taller (> 4 meters in height) and shorter (< 3 meters in height) individuals when compared to those of more moderate heights (3-4 meters). Mitchell et al. (2010, p. 167) speculate that the reason “shorter individuals died out of proportion to their numbers in a population was because they had to compete for browse with other browsers”, consistent with the feeding hypothesis, but that the taller animals died at higher rates “simply because their daily requirement for browse is highest and the amount of browse available to them during the drought almost certainly was well below the amount needed to maintain their metabolism”. It should be noted, however, that the deaths among taller individuals reported by Mitchell et al. (2010) were overwhelmingly male with 87% being male compared to only 13% being female. Given the sex-differences in size in giraffes (Simmons and Scheepers, 1996, p. 782), such an outcome is not so surprising, but it does remind us of an important point. Namely, that while the evidence is not consistent with sexual selection acting on neck length in particular, there are still a number of sexually dimorphic features in giraffes that likely resulted from sexual selection with overall size being the most obvious one.
It's fascinating to see how both the traditional, iconic explanation and the fashionable contrarian explanation seem surprisingly weak given the positive reception they've enjoyed and the influence they've had. They're clean and plausible sounding accounts that seem to have received a fair amount of acceptance despite the lack of convincing supporting evidence for either.
There is an intriguing third hypothesis that could help explain certain observed feeding patterns while still allowing for the neck to be adaptive. It should be made clear that the following is an almost entirely speculative synthesis of the literature and is not being presented as an established hypothesis. Badlagana et al. (2009) raises the important point about how the switch from the ancestral diet to one largely composed of Acacia “could have exposed these populations to toxins like condensed tannins (CTs), which have demonstrable effects on reproductive system function and embryo survival in ungulates”. Furstenburg and Van Hoven (1994), inspired by the apparent lack of correlation between nutritional content and giraffe dietary preference, found that much of giraffe feeding choices can be explained by tannin avoidance. Giraffes have also been observed to tend to feed “from the tops of trees only when new shoots are available” (Mitchell et al., 2010, p. 165), when tannin levels would be the lowest. Given the established reproductively toxic effects of tannins, there may have been strong selection for traits that would allow flexible browsing and so minimize tannin consumption. There is also another possible issue to consider. Acacia, being fabaceous, fixes nitrogen and contains higher levels of protein than most plants (Barnet et al., 1985). Giraffes are foregut fermenters like many other artiodactyls, and high protein diets in other foregut fermenters such as domesticated sheep and cows often induce deadly respiratory side effects (fog fever) when given access to high protein pasture because of conversion of dietary tryptophan to skatole (3-methylindole) by gut bacteria (Hammond et al. 1980). Interestingly, Wood and Weldon (2002) found that much of the well-known ‘giraffe scent’ is attributed to high concentrations of skatole secreted through the skin where it seems to function as an insect and parasite repellent, but such secretions are not found in their closest extant relative, the okapi. It is perhaps possible that the ability to maintain a consistent level of dietary protein afforded by increased neck length was selected on because it enabled continued production of skatole, increasing fitness through a lowering of parasite load and disease risk. Of course, there is very scant evidence for this line of reasoning and it should not be taken seriously absent further work. For now, it's only another nice story.
In the course of writing this post I came across a fascinating and bewildering ethnographic report from the final days of British imperialism in Africa (Cunnison 1958):
I have already mentioned the drink umm nyolokh of giraffe liver and marrow, which many regard as the supreme moment of the expedition. It is said that a person, once he has drunk umm nyolokh, will return to giraffe again and again. Humr, being Mahdists, are strict abstainers and a Humrawi is never drunk (sakran) on liquor or beer. But he uses this word to describe the effects which umm nyolokh has upon him. (It is also used for a man's condition after drinking large quantities of sour milk, which results in a break-down of inhibitions.) I can only assume that there is no intoxicating substance in the drink and that the effect it produces is simply a matter of convention, although it may be brought about subconsciously. Its warmth, its delicious taste and consistency produce an effect of physical contentment on Humr, and probably would do to whoever drank it. It is followed frequently by dreams of giraffe, and I have heard a man wake shortly after drinking it shouting "giraffe on your left". This was regarded as a typical effect. In the waking state, also, men swear they see giraffe through the forest or over the plain where there are none at all. In the absence of any physiological explanation these phenomena may perhaps be regarded as an indication of the extent to which the Humrawi's being is permeated with thoughts of giraffe.
Putting aside the hilarity of "strict abstainers" imbibing large quantities of bovine kumis, there's an interesting possibility of psychoactive indoles becoming concentrated in the liver of giraffes. Skatole is a reaction or two away from being a substituted tryptamine, which also occur in very high quantities in Acacia. While I can find no other sources attesting to such practices, it's a delightful possibility. Cunnison (1958) also writes that a "kind of giraffe mania pervades Humr life". This may be more literal than he intended.
The sexual selection hypothesis appears to be lacking evidence on both anatomical and behavioral grounds. The traditional feeding hypothesis seems to be somewhat weaker than expected when observing usual browsing habits in giraffes, and during times of environmental stress greater height is not linearly associated with greater fitness due to the concomitant increase in metabolic requirements. Still, it's difficult to deny that at least a substantial part of the giraffe’s neck elongation can be attributed to feeding concerns. There are, perhaps, other factors at work such as tannin avoidance or parasite load, but as no directed work has been done in these areas there is little to be said in their favor. Even a century and a half after Darwin, the enigmatic neck of the giraffe still offers questions to be answered and stories to be told.
REFERENCES
Pratt, D. M., & Anderson, V. H. (1985). Giraffe social behaviour. Journal of Natural History, 19(4), 771-781.
Mitchell, G., Van Sittert, S., & Skinner, J. D. (2010). The demography of giraffe deaths in a drought. Transactions of the Royal Society of South Africa, 65(3), 165-168.
Freeman, S. & Herron, J.C. (2004). Evolutionary analysis, 3rd edn. New Jersey: Pearson.
Cunnison, I., 1958. Giraffe hunting among the Humr tribe. Sudan notes and records, 39, pp.49-60.
Furstenburg, D., & Van Hoven, W. (1994). Condensed tannin as anti-defoliate agent against browsing by giraffe (Giraffa camelopardalis) in the Kruger National Park. Comparative Biochemistry and Physiology Part A: Physiology, 107(2), 425-431.
Barnet, Y. M., Catt, P. C., & Hearne, D. H. (1985). Biological nitrogen fixation and root-nodule bacteria (Rhizobium sp. and Bradyrhizobium sp.) in two rehabilitating sand dune areas planted with Acacia spp. Australian Journal of Botany, 33(5), 595-610.
Hammond, A. C., Carlson, J. R., & Breeze, R. G. (1980). Prevention of tryptophan-induced acute bovine pulmonary oedema and emphysema (fog fever). The Veterinary record, 107(14), 322-325.
Wood, W. F., & Weldon, P. J. (2002). The scent of the reticulated giraffe (Giraffa camelopardalis reticulata). Biochemical systematics and ecology, 30(10), 913-917.