How (not) to quantify aurochs-likeness in cattle

I have been trying to quantify aurochs-likeness in extant cattle multiple times and it turned out to be very difficult. It starts with the question of what is a “trait”. For example, is coat colour a trait or is it rather a complex of many traits (what types of melanin are expressed, where they are expressed et cetera)? Some traits also interfere with each other. As an example, leg length (i.e. proportions) influences height and other parameters. If a short-legged bull is only 150 cm tall, it is shorter in height than the European aurochs. If it had the right leg length it might be 20 cm taller, meeting the height of some European aurochs. Is, considering that, both the body size of the animal and the leg length deficient, or is only the leg length deficient? Even more problematic is the thought that not all traits are regulated by the same number of genes. Colour is regulated only by a handful of genes, perhaps roughly a dozen, while the bodily morphology is much more complex and influenced by many genes (I assume so because endocrinology influences the body morphology quite dramatically, and the hormone system is influenced by a very large number of genes). So which cattle individual is more aurochs-like, i.e. closer to the aurochs, one that has the right colour but a totally domestic morphology, or one that has the right morphology but deviant colours? If every defined “trait” (what is a trait?) is assigned an equal value, both individuals would be equally aurochs-like. This neglects the fact that in one individual a lot of genes are “right” and only a handful “wrong” while in the other a lot of genes are “wrong” and only a handful “right”. Therefore, the most objective and useful way to compare the phenotypic aurochs-likeness would be to compare every single gene that has an influence on the phenotype. This is practically near impossible as mammals have between 20.000 and 30.000 genes and we do not know the phenotypic impact of every single allele on every single gene. Therefore, an attempt to quantify even the solely phenotypic similarity to the aurochs necessarily has to be somewhat simplistic. I made lists of aurochs-like traits and scored a number of cattle breeds for those traits, each time a slightly different approach and the result always was that Sayaguesa and Lidia were among the two most aurochs-like breeds, “breeding-back” results were not included. It is very difficult to score “breeding-back” results as they are not as homogeneous as traditional breeds. For example, the horn size of Heck cattle varies greatly between individuals, herds and lineages. Averaging the traits found in the “breeding-back” results each would not account for the variability found within them. What can be done is scoring single individuals of “breeding-back” breeds/populations, but that leaves a lot of room for bias. One can take the best or the worst individual, or trying to figure out which individual might be representative of what is average for the breed, but that is very difficult and would, again, not account for the variability in the populations. 

All in all, I think trying to systematically objectively quantify the aurochs-likeness of cattle can only provide a very rough impression that is not any more useful than just looking at the cattle and comparing them. 

 

How bulky was the aurochs?

As mostly only bones are preserved of the aurochs, reconstructing the soft tissue surrounding them necessarily involves speculation. The anatomy of living wild bovines and cattle provide useful clues for that. But they do not tell us how pronounced the muscles really were, how large the intestinum was (this is relevant as domestic cattle have a much larger intestinum than wild bovines) and how much tissue surrounded the bones in general. 

Looking at a skeleton, there often is the tendency to underestimate the bulk of tissue that surrounded the bones. Those who are familiar with dinosaur paleoart will know of the term “shrink-wrapped” dinosaurs. That term arose when it was recognized that most dinosaur reconstructions from the 1990s and early 2000s were too skinny, often to a degree that makes anatomically no sense (which was a result of a countertrend to making dinosaurs extremely bulky in earlier 20^th century and 19^th century reconstructions). Nowadays dinosaur reconstructions much more appreciate the musculature those animals most likely had. I was not shrink-wrapping my dinosaurs in my drawings already back in the time before the term was even used in dinosaur paleoart. So, it is a bit ironic that I tend to make the same anatomical mistake in my aurochs reconstructions. 

The bones were surrounded by musculature, which has the greatest impact on the outer appearance of the animal. But the musculature was surrounded by fat tissue, which added a few millimeters to centimeters, by the skin, which added a few millimeters to centimeters, and by hair, which also added a few millimeters to centimeters. This necessarily involves a fair bit of guesswork to the reconstruction. I am always, without exception, unsure how much bulk I should add to the skeleton when doing a reconstruction. The actual shape of the animal would also not always be the same when fully grown, but it changed during the season (fatter during fall, skinnier during the end of the winter) and also across its lifetime as aging individuals get heavier, the bulls in particular. Comparing my recent reconstruction based on manipulating a photo of the Taurus bull Darth Vader III to my model of the Sassenberg bull (go here), I came to the conclusion that my model is probably not accurate. It is too skinny, particularly on the legs and the abdomen, making it a bit difficult to imagine this model as a living being. So I made some anatomical sketches trying to better appreciate the bulk of the soft tissue that surrounded the bones. 

 

The trunk 

The trunk is particularly difficult because even with seeing the ribcage and the curve of the spines in real, it cannot be derived from the skeleton how large the intestinum was. Domestic cattle have very large intestina, giving them the heavy appearance they have. Wild bovines, on the other hand, usually have a waist that narrows caudally (instead of being the centre of the mass). For my sketch, I tracked out a photo of the Store-damme skeleton that I had corrected anatomically using GIMP. The blue line shows what a young but fully grown aurochs might have looked like, similar to young fighting bulls and young wisents. The green line shows what an old bull might have looked like, based on the waist anatomy of old wisent bulls. The red line shows the domestic condition that is exhibited by most domestic cattle breeds. I have no precise anatomical method to come to these lines, I am looking at bovines in flesh and blood, guessing and drawing what I consider anatomically plausible. I am only quite confident that the red line is not plausible for an aurochs, since it is the domestic condition not found in any wild bovine. It also looks quite weird on a trunk as short as in the aurochs. 

 

The head 

 

The outer shape of the head also changes quite with age. Young bulls have a slender head while older bulls are bulkier. That’s why my 2019 model looked a bit juvenile (or actually subadult), its head was too slender. I wanted to do a model more credible for a grown bull this time, but it seems that I still “shrink-wrapped” it. Here are sketches that were done tracking out the London skull in frontal view: 

 

The left one is a “shrink-wrapped” version, the left one a version that I consider more plausible. Again I had no particular anatomical method to come to this sketch, but I tried to appreciate the facial muscles that undoubtedly were there and also the skin and fat tissue. 

The other sketches are based on the skull of the Sassenberg bull: 

 

The sketch on the right shows a “shrink-wrapped”, not anatomically plausible, version. It has a paper-thin skin, barely any facial muscles or salivary glands (that sit behind the lower jaw muscles), and there is barely place for the trachea and the oesophagus. The sketch in the centre is anatomically plausible to me, perhaps for a bull in its prime. The left sketch shows the maximum bulkiness that I still consider anatomically plausible. Old bulls might have looked like that. Looking at my most recent model (here) I think it is closest to the shrink-wrapped version, thus not anatomically plausible. So I added some bulk to the head and neck region with GIMP (and also painted an eye). The result looks much more plausible and life-like, I think: 

 

This shows that some aspects of the aurochs’ life appearance are very hard to strictly derive from the skeleton when trying to be very precise, and that trying to fabricate the “perfect aurochs model” is a continuous learning process. My next model is going to be better. I think I will reconstruct the Vig bull and Cambridge specimen (? cow) next time. 

 

Turning Darth Vader III into an aurochs

I have attempted reconstructing the aurochs by photo manipulation using GIMP on photos of aurochs-like cattle multiple times, but this time I think I might have fabricated the most accurate one I have attempted so far. 

 

I used a photo of Darth Vader III, a Taurus bull that is a son of Londo (F2 Sayaguesa x (Heck x Chianina)) and Laniana (which had a little bit of Lidia in her genealogy), thus is 1/32 Lidia, that I took in the Lippeaue in 2022. I superimposed a photo of the Store-damme aurochs skeleton, which I had previously corrected into an anatomically correct posture on the photo. Then I started to manipulate the proportions and shape of the bull according to the anatomically correctly positioned skeleton. I increased the length of the legs, changed the hump by enlarging it slightly, reduced the length of the dewlap, made the neck slightly longer. For the head, I used a photo of the skull from the Sassenberg bull. I made the snout slightly longer, and the postorbital region of the head. Then I made the bodily morphology more like a wild bovine by reducing the bulk of the waist and belly. All the other aspects of the original bull photo fitted the aurochs skeleton. Honestly, I was surprised that I did not have to change the bull that dramatically. The horns of the original had to be removed with the stamp tool, I replaced it with a horn from the Heck cow Erni of the Steinberg/Wörth lineage that I skewed slightly.

This is the original photo of Darth Vader III: 

 

This is the result: 

 

I am very happy with it. It looks rather plausible and credible for a real (wild) bovine to me, more so than my recent model which is based on the same skeleton and skull: https://breedingback.blogspot.com/2023/04/life-reconstruction-model-of-sassenberg.html 

The photo manipulation shows me the deficiencies of the model, my next one is going to be better. 

 

I think that less-derived Lidia, Spanish fighting cattle, are the cattle that come closest to this reconstruction by far. In my opinion, this phenotype is roughly achievable in “breeding-back”, as long as wild yak, Java banteng and Lidia are in the mix and used wisely. 

Why Equus ferus should not be used for the wild horse

The use of the binominal name Equus ferus for the wild horse is extremely common, and I used it myself in the past. However, this binomen is problematic and should not be used except in a list of synonyms of Equus caballus. 

My readers will know that I prefer to use Bos primigenius for the aurochs, thus the later-published binomen that is frequently used for the wildtype, while I treat the species of the horse (including the wild horse and its domesticated derivatives) under the name Equus caballus, which is usually used to refer to the domestic form only. This may seem inconsistent at first, but examining the problem at a deeper taxonomical level shows that it actually is not. 

I prefer Bos primigenius because this binomen has a holotype that is undoubtedly a wild specimen, the Haßleben aurochs skeleton, while Bos taurus lacks a holotype or lectotype. Type specimen are important because they clarify which species is referred to by the name. Which species is Bos primigenius? The species that the Haßleben skeleton belonged to, whatever a species definition you work with, no matter how narrow or wide your species definition is. Thus, the lack of a type specimen is problematic because it makes the situation far more ambiguous. 

As for the horse, neither Equus caballus nor Equus ferus have a type specimen. For Equus caballus, this is less problematic because it referred to domestic horses, while for Equus ferus it is a big problem because of synonymity. If two binominal names are synonymous, i.e. they refer to the same species, the rule of priority of the ICZN dictates that the name that was published first is the one to be used (the senior synonym), while the younger one is the junior synonym (as a side note, synonymity does not make the junior synonym invalid [as long as it is described accordingly to the rules of the ICZN], as it is often believed, it just means that it does not have priority and shall not be used). In the case of domestic animals, where the binomen that was erected for the wildtype exclusively often is the junior synonym, opinion 2027 of the ICZN allowed the use of the names for the wildtypes for the entire species. This includes Bos primigenius, but also Equus ferus. The ICZN explicitly assumed in opinion 2027 that Equus ferus is based on the ancestor of domestic horses, which they call “tarpan” in their document. For Bos primigenius, there is no problem following opinion 2027 because it is based on a specimen that was indeed a member of the wildtype of cattle, for Equus ferus there is indeed a problem. In order to see if Equus ferus is justifiable as the binomen of the wild progenitor of the domestic horse, we have to look at the original description of this name. 

I used Equus ferus in the past because, like everyone else, I did not check the original source and assumed that everything is alright with that name. This is the reason why it is that widely used as the name for the wild horse. For some reason, I assumed Equus ferus was based on fossil or subfossil remains of definite wild horses in Europe. When I checked the original source, I realized that this name should not be used for the wild horse. 

Equus ferus is introduced by Pieter Boddaert in his 1785 work Elenchus animalum. Boddaert gives a short description mentioning long ears, a short curly mane, a blackish grey mouse-coloured body and a short tail, and refers to the written description of alleged wild horses encountered by S. G. Gmelin in Voronezh in 1769. That is the original description of Equus ferus. The big problem is that it is far from clear what those horses spotted by Gmelin were. Whether the horses were wild, feral or hybrids of wild and feral horses. The description fits what other writers reported of free-roaming horses at that time and region. At least one member of the herd was a black domestic mare, as reported by Gmelin himself. It is generally unclear what the horses of that time and region were, and there never was a consensus among authors, be it contemporaneous ones or after the extinction of those populations in the 19^thcentury to this day. And in the lack of a type specimen, be it a holotype or a lectotype, it cannot be ascertained anymore. This case shows why type specimen are important. 

Therefore, there is no certainty that Equus ferus is based on a wild representative of the species that includes the domestic horse. If domestic animals and their wildtypes are regarded as separate species, the name to be used for the species of the wild horse would be Equus przewalskii because this binomen was the first one to be erected based on an undoubtedly wild specimen of the species. It has both a holotype and paratype which were wild Przewalski’s horses. I usually treat wild and domestic animals as members of the same species, and the first name erected for this species is undoubtedly Equus caballus by Linnaeus 1758. Linnaeus also erected a senior synonym for Bos primigenius, namely Bos taurus. However, in this case I prefer to follow opinion 2027 because the junior synonym conserved is based on a type specimen that was undoubtedly wild, while Equus ferus is based on a description which itself is based on a description of an encounter with horses of unknown status. Therefore, following opinion 2027 in the case of Equus ferus is not useful. For the same reason Equus ferus should not be used on subspecies level for the subspecies that was ancestral to the domestic horse. There is not the slightest bit of scientific evidence that the horses encountered by Gmelin in 1769 were members of the form that was domesticated 5000 years ago. For the same reason Equus ferus should not be used for Pleistocene wild horses. It should not be used for wild horses at all. If one wants to use Equus ferus solely to refer to the horses historically called “tarpan”, it is questionable if these populations deserve a subspecies status as it is unclear what those horses were. So neither Equus ferus nor Equus ferus ferus are taxonomically justified. 

This of course leaves the question what the subspecies that was the predecessor of the domestic horse should be called then. For this question to be answered, a rigorous assessment of the Holocene wild horse material that has been found in Eastern Europe would be necessary. It is well possible that someone already erected a taxon based on that material that would have priority. Whether or not a Pleistocene wild horse name (and there are a lot of those) is applicable to this form remains to be seen, this would have to be morphologically and genetically evaluated. When a subspecies is not yet scientifically described, it is referred to as species X ssp., in the case of the wild horse that was ancestral to the domestic horse it would be Equus caballus ssp. – a possible way to “rescue” Equus ferus is to assign undoubted Holocene European wild horse material to Equus ferus as a lectotype. As long as this has not been done formally, Equus ferus should not be used as the binominal name of the wild horse. 

 

The lateral horn orientation

When it comes to the horns of aurochs and cattle, I usually differentiate them into three factors to compare them: the actual shape of the horn, the dimensions (length and diameter) and the orientation relative to the skull. All those three factors can vary among cattle. When thinking of the horn orientation, I only considered the orientation relative to the snout. In the European aurochs, this angle usually varied from 50-80°, with some individuals being exceptions. I did not notice that the horn orientation also has another variable that has quite an impact on how we perceive the horn “shape” (the shape actually does not change, but it appears as if it did when the orientation is changed), what I call the lateral horn orientation. This factor is the orientation of the horn in the plane that is 90° to the sagittal plane of the animal, in other words, the horn orientation relative to the side of the animal. I did not realize the lateral horn orientation is a factor before I did a photo manipulation morphing the Taurus cow Lerida into an aurochs cow, where I had to slightly elevate the lateral horn orientation of the cow in order to make the horns appear more aurochs-like. 

Another good example that the LHO is relevant are the horns of many Sayaguesa. Looking at the head of one of the Sayaguesa cows grazing in the Lippeaue, at first glance one would say the horn tips do not face inwards enough because they point right towards the sky and that the primigenius spiral is not present: 

But elevating the LHO, so that the proximal half of the horns do not point downwards but slightly upwards, an aurochs-like horn curvature with the primigenius spiral appears (the shape of the horn was not changed, showing that it already had an aurochs-like shape): 

So I did a sketch playing around with the LHO: 

 

I sketched an aurochs horn that I copied and mirrored, and positioned with a different LHO four times. The uppermost sketch shows a LHO rather elevated, as can be seen in the skull fragment of Asti and Gramsbergen (both likely belonged to cows). The second one counting from the top shows the LHO slightly elevated, as exhibited by the vast majority of aurochs skulls (f.e. the Sassenberg bull, the Nagybajom bull etc.). The third one shows the LHO in a horizontal position, seen in for example the Arezzo skull, some Lidia, some Maronesa, some Sayaguesa bulls. The fourth shows the LHO oriented downwards, a condition that did not appear in the wildtype (at least no skull with such a horn orientation has been found so far), but is found in many Lidia and many Sayaguesa. 

 

Looking at the horns of the, in my opinion, most aurochs-like “breeding-back” population, the Taurus cattle from the Lippeaue, it turns out that their LHO could be more elevated, as seen in the Lerida example. Watussi might improve this a bit. The Hungarian Taurus population from Hortobagyi, which has influence from Watussi, indeed has a more elevated LHO on average, but there is also Steinberg/Wörth Heck cattle and Hungarian Grey cattle in the mix. Wild yaks would help as well, perhaps better than Watussi as their horns are identical to those of the European aurochs. 

 

Crossbreeding proposal: Chianina-Watussi

Chianina are famous for their large body size with sometimes long legs, Watussi are for their very large and thick horns. Now we know that the original aurochs phenotype was very large with mostly large to very large horns (go here for a post on particularly large-horned aurochs). I think the combination Chinaina-Watussi bears potential for achieving cattle that are both large and large-horned and even more than that. 

The problem is that Chianina is large but very small-horned, and Watussi is large-horned but small. Both horn size and body size are quantitative traits which are controlled by many different genes that all have a more or less large or small impact on the final phenotype. That means that if one wants the result of cattle that have the Chianina body size but not horn size, and the Watussi horn size (or slightly smaller) but not body size, the breeding must use the second Mendelian law, otherwise the breeding takes forever. And most likely a larger number of individuals is required, not only a couple of individuals but rather herds of 20-30 (which happens to be the natural Bos primigenius herd size). So at first, a herd of Watussi x Chianina should be produced, as many individuals as possible. The phenotype of the F1 individuals is completely irrelevant as it is the result of maximum heterozygosity, thus combinations of alleles that will not be present anymore as the population becomes genetically stabilized. Both horn size and body size will be somewhat intermediary between the parental breeds, but that is not really relevant yet. It’s the second generation, the F2, where the phenotype of the animals becomes relevant. So a herd of F2 Watussi x Chianina should be produced. Some will have an indicine hump (probably), some will not. Some will have smaller horns and smaller body size, some will have large horns but small body size et cetera. As the focus is on creating a population of large cattle with large horns, only those F2 that have large horns and body size should be chosen for further breeding. 

It is even possible that the combination Watussi-Chianina has potential for a correct wildtype colour with at least some degree of sexual dichromatism. This may sound counterintuitive when two breeds are crossed that both have no visible sexual dichromatism and one is completely white and one is homogeneously reddish-brown. But one should consider the genetics of the two breeds. Chianina has the wildtype allele on the Extension locus, but the colour is diluted by at least two dilution alleles. Watussi does not have these dilution genes, otherwise it would not have that strong expression of red pigment – it has a mutation on the Extension locus that disables the production of black pigment. So a F2 combination of both has the potential for individuals that have the Extension wildtype allele and the production of both black and red pigment enabled. A hint for that is the colour phenotype of the Watussi x Chianina cow of the Auerrind project: 

 

© Claus Kropp 

It clearly has the wildtype Extension allele expressed (it is dominant over the recessive mutation disabling black pigment) and the colour is barely diluted if at all (possibly because some of the dilution alleles of Chianina are recessive). So a F2 of this combination could potentially result in, among other colours, wildtype colour that is not diluted. Sexual dichromatism is possible as well because Chianina might have some degree of sexual dichromatism or even a very strongly expressed one, masked beneath the dilution factors resulting in a white colour in both sexes. That possibility is there since some Sayaguesa x Chianina bulls produced in the Lippeaue had a perfectly black wildtype colour while the cows are always lighter in colour. Therefore, there is at least the chance to achieve aurochs colour from two breeds which do not have aurochs colour at all, because of the genetic background of the colours of the parental breeds. 

In the F2 Chianina x Watussi generation, I would pick those that are large with large horns for further breeding. Some of them might have indicine traits such as the fleshy indicine hump or hanging ears, and the body morphology might not be completely taurine because of the more or less high portion from Watussi. That is why I would backcross the F2 with good Chianina again. And, as you can surely guess by now, create an F2 of this backcross generation. Creating true F2s is the most important step for genetic reasons. The backcross generation might look quite good already, but it will have a full set of Chianina chromosomes. That means they are heterozygous for all the optically good traits and will pass on the alleles for small horns and various colour dilutions. By using the second Mendelian law that applies to a true F2 generation, these alleles can be selected out much more easily than by conventional breeding. So, in the F2 of the Chianina x F2 (Chianina x Watussi), I would only continue to breed with those individuals that are large, have a taurine morphology and large horns, and – as far as possible – also a correct colour. From this selection event on, I would continue to breed with the suitable individuals from that combination in a conventional fashion otherwise the gene pool might become too narrow. It is even possible that not only body size, horn size, colour and morphology of this population is good but also the horn shape, as Chianina often have forwards-facing horns with sometimes a useful curvature. 

 

This scenario takes four generations (8-10 years) and might result in very satisfying individuals that will pass on the desired traits more or less reliably. 

 

 

Would it be crazy to use hybridization in "breeding-back"?

As my readers will know, I suggested cautious hybridization with extant wild bovines that would add traits hardly achievable using domestic cattle only in “breeding-back” in a couple of my posts, after I initially opposed that idea. In this one for example, I suggest using wild yaks in “breeding-back”. But the Java banteng would add value as well, and using both in cautious doses would be add all of the traits that are difficult to achieve with domestic cattle only. The kouprey, which would add all the desired traits both the wild yak and banteng would add and has a slightly more aurochs-like morphology, unfortunately is not available for hybridizing as it is most likely extinct. To give a quick summary of the benefits the wild yak and the Java banteng would add to “breeding-back”:

Wild yak: 

- Very large, up to 205 cm withers height in bulls what equals the large size of European aurochs; difficult to achieve with domestic cattle only, where the upper height limit might be roughly 180 cm; well-marked size dimorphism between the sexes 

- Perfectly aurochs-like horns in both sexes. The horns of the wild yak are really identical to those of the European aurochs in terms of size, curvature, orientation to the skull and even colour. They have a very pronounced inwards-curve which is lacking in most aurochs-like cattle 

- They have a large shoulder hump and a long snout, two traits which are rarely as expressed as in the aurochs in domestic cattle 

- The physiological fitness is higher than in domestic yak and most likely domestic cattle 

- The food choice overlaps with that of cattle/aurochs, they are predominantly grazers 

Java banteng:

- A very well-marked colour difference between the sexes that is always present. This is extremely rare in domestic cattle as domestication reduces sexual dimorphism

- Aurochs-like body size (up to 190 cm which is not quite as large as wild yaks but still significantly larger than most domestic cattle) 

- Roughly aurochs-like morphology (hump, short trunk) 

- It ecologically overlaps with cattle/aurochs, although they browse more, including fruits 

 

Cautious hybridization and wise selection might thus speed up the process of achieving an aurochs-like animal significantly and might have a better end result. But there are, as one can imagine, objections against hybridizing with other species, which I shared until recently. Those objections are two main ones: 

 

Argument #1. A different species would add many undesired, non-aurochs traits, some of which are practically invisible and thus hard to select against, which might distort the authenticity and ecologic suitability of the results. 

There is valid truth to that argument. For example, wild yaks are adapted to higher altitudes than the aurochs, which was a lowland animal, and have larger lungs. They are also adapted to more severe cold, having fewer sweat glands and a different pelage. Also, there are skeletal differences, such as a different vertebrae count. Wild yaks also browse less than cattle/aurochs as they are more strict grazers. Banteng are adapted to a much hotter climate and browse more than cattle. There are also behavioural differences in both species. Yaks, for instance, have a different aggression display behaviour by presenting their large tail tuft, while banteng rely more on display in intraspecific competition than fighting head to head like cattle/aurochs and wild yaks. 

However, it is not impossible to get rid of these undesired behavioural, ecological and morphological traits. The solution might be absorptive breeding while running a strict selection regime. That means that each hybrid generation would be backcrossed with taurine cattle, but only those hybrids that still have retained the desired traits added by the species would be kept for further breeding. After eight generations, for example, the genealogical portion from the other species would be as low as 0,3%, therefore there would be a high chance that the undesired traits are washed out while the desired ones remain in the population when selected appropriately. 

 

Argument #2. “Breeding-back” cattle with a tiny but important bit of introgression from wild yak and Java banteng would not be pure domesticated aurochs anymore. Thus, the argument that “breeding-back” cattle are a native species in European nature would be gone, as they would more or less be hybrids. 

There are two problems with this objection. On a more general note, introgression between closely related species at some point of time has been found everywhere it was looked for in mammals, including our own species Homo sapiens, so it seems that hybridization (or more precisely, introgression) is a very widespread phenomenon in the process of speciation and evolution in general. In the case of wild bovines, Cambodian banteng have introgression from the kouprey, wisent have homeopathic introgression from aurochs and there may be more cases which we are currently not aware of. So introgression per se is not “unnatural”, quite the contrary. Furthermore, it is not technically true that domestic cattle are “pure” albeit domesticated aurochs. Domestication itself is an anthropogenic process, resulting in organisms that differ from the wildtypes that were shaped by nonhuman evolution. A wild aurochs’ genome is 100% wild aurochs. A domestic cattle genome is 100 - x % wild aurochs, x being the portion of the genome that is made up by alleles that mutated after domestication. I do not even dare to speculate how large or small that portion is, and it is certainly different from breed to breed to a certain degree. Thus, domestic cattle are not “pure” aurochs, but rather partly wild aurochs and partly a domestic mutant. So, the originality of the aurochs is gone as a result of domestication anyway. If another species is bred in, the non-aurochs portion would increase dramatically in the first couple of generations, but decrease in the advanced generations as a result of absorptive breeding. It is even possible that, when selected wisely, the aurochs portion could actually increase by introgression from wild yak and banteng. I say that because I consider it likely that, as long as the aurochs-like traits in wild yak and banteng are not the result of convergent evolution (which I do not consider particularly likely), these traits are caused by the same wildtype alleles as in the aurochs. For example, the identical horn curvature between aurochs and wild yak might be caused by alleles shared from a common ancestor. Introgression might reintroduce these alleles into the domestic cattle population. The same could be the case for the sexual dichromatism in the Java banteng. In this case, introgression would reintroduce wildtype alleles found in the aurochs which are absent or at least very rare in domestic cattle. So the “genetic purity” argument is not convincing to me. Apart from that, the introgression might also introduce physiological fitness-related alleles from banteng and wild yak which are superior to those of domestic cattle in terms of physiological fitness, thus providing an advantage for “breeding-back” cattle compared to the domestication-induced disadvantage that domestic cattle usually might have. 

 

Therefore, I think small doses of controlled hybridization/introgression from the two wild bovines most suitable, and subsequent wise selection, is not a crazy idea at all and could be very beneficial for the ecologic and phenotypic goal of “breeding-back”. What would be crazy, on the other hand, would be hybridization without a plan and without strict selection. 

 

 

10 years Breeding-back Blog + the Gramsbergen aurochs

Since May this year, I have been running the Breeding-back blog for ten years now. So far, it had more than 1,2 million page views, 500 posts and more than 2000 comments. I am very happy with that and want to celebrate it by presenting my recently finished aurochs clay model, based on the Gramsbergen aurochs. 

 

What I call the Gramsbergen aurochs is a specimen that is known only from a skull fragment with horn cores that is on display at the Museum of Gramsbergen. I know two photos of this skull fragment, that are unfortunately not online anymore. One shows it in frontal view, one in dorsal view. What is interesting is that the lateral horn orientation of that specimen is comparably vertical, around 45° in frontal view. It is very hard to tell if the skull fragment is from a bull or a cow, but since the orbital bosses are not much protruding, the postorbital region does not look very massive and the horn shape does not resemble most of the definite bull skulls, I assume it is more likely that this fragment belonged to a cow. 

For my model, I used the Sassenberg skeleton as a reference for the cranial and postcranial morphology since it is the only complete definite cow skeleton from Europe that I am aware of. As always, the model was done taking measurements from the photos that I have access to, to ensure maximum accuracy. 

For the colour, I chose a rather standard scheme out of the many possible shades that an aurochs cow might have displayed. Some depictions in Lascaux show exactly that colour. 

The head itself resembles that of Sayaguesa and Lidia cows or a mix of both. I would imagine that a combination of Sayaguesa, Lidia, Watussi and Maronesa could result in an animal that resembles this particular individual quite closely at least concerning head, horns and colour. Morphology is more difficult to achieve. 

 

Post #500: How to speed up "breeding-back"

Selective breeding with domestic cattle cannot, under any circumstances, revive the aurochs. What can be achieved is what I call the greatest similarity to the aurochs that is achievable using domestic cattle. I prefer to speak of reaching that instead of having “bred-back” the aurochs or having “completed” breeding back (a process that is actually never completed). The animals will remain domestic cattle, but with a high degree of morphological/phenotypical similarity and ecologic equivalence to the aurochs and with very similar if not identical social behaviour patterns. Some domestic traits, such as a reduced brain volume or sexual dimorphism, or a not quite as athletic bodily morphology et cetera will remain, but an identical colour scheme with a well-marked sexual dichromatism (which is not the same but part of the overall sexual dimorphism), the same body size range, a very similar if not identical horn shape and size, aurochs-like proportions, a well-expressed hump, an at least roughly similar skull shape and of course the ecologic and behavioural similarities can be achieved by breeding with domestic cattle. These cattle would be the prime option for replacing the aurochs ecologically in the wilderness. 

As readers of my blog will know, breeding for the greatest achievable similarity to the aurochs is a long-term endeavour. That is because cattle are a comparably slowly reproducing species (nine months of gestation, roughly two years until sexual maturity, litter size of only one individual) and most breeding projects do not want to sacrifice genetic diversity by using a high degree of inbreeding (which would speed up the process but has its downsides to health and adaptability). How could this long-term process be sped up? 

Concerning the greatest achievable similarity to the aurochs, there are many ways to the goal. Many combinations of cattle breeds might deliver all the desired traits (morphologically as much as ecologically), several projects are running at the moment and one can always start from scratch. But is there a most efficient breed combination, one that includes all desired traits but has the least amount of undesired traits? How to breed more efficiently than the current breeding methods employed by the projects that are running at the moment? I see three possible scenarios: 

 

-) Starting from anew with the most efficient breed combination. I have been puzzling about what might be the most efficient breed combination, and I think it would be a set of five breeds: Lidia (body morphology, sometimes also horn shape, pelage/forelocks), Chianina (body size, sometimes horn orientation and shape and sometimes leg length and morphology), Maronesa (perfect wildtype colour, sexual dichromatism, useful elements of horn curvature), Watussi (horn volume), indicine breed like Deoni (very short trunk and long legs, very small udder). The most efficient breeding scheme would be to start with a herd of Chianina cows that are covered by a good Maronesa bull, and producing a herd of F2 of that combination, and parallel to that a F2 of Lidia x Watussi and Lidia x Deoni. Ideally, there would be three lineages: one that is large, has the right colour + dichromatism, aurochs-like horn curvature and a good morphology; one that has the Lidia-like morphology with an aurochs-like horn volume and possibly shape; and one that has the Lidia-like morphology with a very short trunk and long legs. In the next step (third generation), I would create F3 of the Maronesa-Chianina lineage to stabilize the good traits plus F1 of the Lidia-Watussi and Lidia-Deoni F2. F1 will not be stable in inheritance, so creating an F2 from this new combination would be required, and only the resulting F2 with a Lidia-like morphology, short trunk and long legs, and aurochs-like horn volume with a curvature that is useful should be kept for further breeding. These F2 should then be bred to the F3s of the Maronesa-Chianina lineage, and the resulting F1 should be used to create F2 of that combination. If some of the F2s of this combination has all the desired traits, they can be used for further breeding with less strict selection. 

Minimum generations required: 6 generations (roughly 15-20 years)

Advantages: the result could be very aurochs-like 

Disadvantages: requires a large number of parental individuals to avoid too much inbreeding and relying too much on luck. Deoni would probably have to be imported from Asia, which might be impossible for practical reasons. 

 

-) Strong selection within the most aurochs-like currently available “breeding-back” population. The most aurochs-like currently available “breeding-back” population would be, in my opinion, the Lippeaue Taurus cattle population. What is lacking, to a certain degree, is horn volume and short trunk length (the latter is a problem that most taurine cattle have). Because of that, an ideal combination would be a herd of large Lippeaue cows (like Bionade, Larissa, Linea) plus a Watussi-influenced bull from Hortobagyi or from the Auerrind project that has a truly large horn volume. The resulting F1 offspring from this combination would then be selected – as the P generation is less genetically homogeneous than purebred P, the resulting F1 would not be homogeneous either, so selection is necessary. With the good F1, F2 would be created and so forth, and the animals of each generation are selected and only those that are as good or better than their parents are used for continuing the breeding. 

Minimum generations required: Indeterminable, as it depends on how good the Fs from the respective generations are, which in turn depends on luck and the number of animals used. 

Advantages: Would be the least effortful variant as it would be based solely on cattle that are already used in the project or projects. 

Disadvantages: Using current “breeding-back” cattle only might go at the expense of certain traits, mainly short trunk length. A suitable indicine breed like Deoni might have to be crossed-in, what will require more generations.  

 

-) Using introgression from wild yaks and Java banteng on an already aurochs-like population. Wild yaks would add perfectly aurochs-like horns which are found only very rarely in domestic cattle, very large size (up to 205cm withers height in bulls), long snouts, large hump, cold tolerance and physiological fitness. Java banteng would add a well-marked sexual dichromatism that is very rare in domestic cattle, large body size (up 190 cm withers height in bulls) with a well-marked size dimorphism and a short trunk. So both species would help to achieve an aurochs-like extent in aurochs traits that are rarely found in domestic cattle. I would start with two herds of Lippeaue Taurus cows, which are suitable for two reasons: a) they are the most aurochs-like “breeding-back” cattle to date, b) their comparably large body size should lead to no problems during gestation and calving. I would inseminate one herd with wild yak semen, and one herd with Java banteng semen. Of the resulting hybrids, the males would be infertile. For the herds with the hybrid cows, good Taurus bulls would be chosen. The resulting backcrosses would be selected: those that have the desired traits added by Java banteng and wild yak would be kept for further breeding, those that do not would not. The B1 individuals would be backcrossed with the Taurus bulls once again. The resulting B2 would be selected like the B1 generation. The B2 bull with the most desirable (= the desired traits added by Java banteng/wild yak present, as few as possible other traits added by these species present) would be chosen as sire for the good B1 and B2 cows. The resulting fourth-generation animals would be selected after the same principle. The best (= those that have the least influence from wild yak and Java banteng but still have the desired traits) of these two lineages would be kept for further breeding. The fifth generation animals would be F1 from the banteng lineage x wild yak lineage. As you can guess by now, F2 from this combination would be created. Only those that have the desired traits added by banteng and yak, respectively, would be kept for further breeding. They can be backcrossed with Taurus endlessly to wash out the undesired influence from the two species, as long as those individuals that have the desired traits added by them are kept for further breeding. The result would be very aurochs-like in having “perfect” horns, being very large (possibly larger than un-hybridized “breeding-back” cattle) and a more well-marked dichromatism than would be possible without hybridization. Also, they would have long snout as in the aurochs and a well-developed hump, possibly more so than “breeding-back” cattle without hybridization. 

Minimum generations required: 6 to theoretically endless (20 years to endless) 

Advantages: some important traits might be achieved better and faster than without the hybridization from wild yak and Java banteng, that is body size, hump size, udder size (or lack thereof), snout length, horn curvature and sexual dimorphism. 

Disadvantages: the hybridization with another species is controversial, some of the traits added by them are invisible or behavioural but have an impact and can only be bred out by a practice of backcrossing and selecting as much as possible, which requires several generations at least. 

 

As you see, none of these methods is entirely unproblematic, and none is noticeably faster than the other. Does this mean we would have to wait at least for another 20 years until we see a “complete” result in “breeding-back”, an individual that has all the desired and achievable traits? Probably. Some individuals, like a 25% yak or banteng individual, or a 25% Watussi 75% Lippeaue Taurus might look quite satisfying in a much nearer future already, although they would be imperfect from the breeding perspective (not genetically stable yet, too many behavioural, morphological and ecological traits from the hybrid parents, lacking some traits et cetera). Highly ambitious “breeding-back” with cattle simply takes a while. However, to be honest, I think that some herds, like the Lippeaue herd (perhaps combined with Auerrind or Hungarian Taurus cattle for horn volume), are already aurochs-like enough to be released in the wild, basically in a dedomestication experiment (that is, releasing the cattle in a reserve and not interfere with the natural breeding of the cattle) and let natural selection do the rest while the cattle are doing their ecologic job in a rewilding area. It must be noted, however, that the dedomestication process takes even longer than artificial selection. Having made this list, I think that using small and wise doses of hybridization with wild yaks and Java banteng is the most tempting of the three scenarios, as it would result in the most impressing and perhaps most aurochs-like cattle. An upcoming post is going to examine if using hybridization is a crazy idea. 

 

Did the African aurochs die out because of aridification?

It is well-known that the aurochs was wiped out by humans. It was mainly a combination of hunting, competition for habitat and competition with livestock that caused the extinction of this wild bovine. This goes for the wildtype as a whole, but it does not rule out that single populations or subspecies died out because of “natural” (i.e. non-anthropogenic) causes. This was, for example, the case with the Sicilian dwarf aurochs, Bos primigenius siciliae, which likely was outcompeted by the large mainland form when a land bridge between the mainland and the island formed at the last glacial maximum. For the African aurochs, Bos primigenius mauretanicus, there are not many works that cover possible extinction causes for this form specifically. However, there is at least one, which is Faith 2013 [1]. This work attributes the extinction of the aurochs in Africa to an aridification of its environment in North Africa around 6kya. How plausible is this? 

This idea sounds plausible for three reasons: 1) the Holocene was anything but climatically stable, and during the Atlantic period there was the so-called Holocene Climatic Optimum, during which it was warmer than today and even the Fertile crescent aridificated [2], 2) the aurochs was an animal that preferred wetter habitats and did not inhabit steppe, semi-deserts or deserts, thus an aridification could have been harmful, 3) the Mediterranean see would have hindered the aurochs from migrating northwards to less arid and more humid regions. 

There is, however, a significant problem with the idea that aridification caused the aurochs to disappear in Africa: the African aurochs did not disappear at 6kya. Bone findings suggest that the African aurochs survived at least until 1000 BC [3,4]. Furthermore, the hunting of aurochs is documented in ancient Egyptian art and hieroglyphs [3]. Less unambiguous is the mentioning of bovines with horns so large and forwards-facing that the animals had to graze backwards in Egypt by Herodotos from the 5^th century BC [4]. 

I would not be surprised if the Nile served as a refuge of the aurochs during the aridification in the Mid-Holocene, as the aurochs is dependent on wet and plane areas. After the warm and dry period during the Holocene Climatic Optimum was over, the range of the aurochs would have expanded again. But now the bovine had a competitor, humans, who competed for exactly the same habitat that the aurochs preferred. Additionally, humans hunted the bovine. So it seems that in North Africa exactly the same happened as in Europe and elsewhere, the aurochs simply had no chance in competing with humans for the same habitat additionally to being hunted. The aridification as a result of the changing climate during the Mid-Holocene cannot be blamed for the extinction of the African aurochs. 

 

References

 

[1] Faith, J.T.: Late Pleistocene and Holocene mammal extinctions on continental Africa. 2013.

[2] Vinos, J.: Climate of the past, present and future: A scientific debate, 2^nd ed. 2022.

[3] Beierkuhnlein, C.: Bos primigenius in Ancient Egyptian art – historical evidence for the continuity of occurrence and ecology of an extinct key species. 2015. 

[4] van Vuure, C.: Retracing the aurochs – history, morphology and ecology of an extinct wild ox. 2005. 

 

New Auerrind breeding herd has been assembled

A few days ago, the Freilichtmuseum Lauresham has published a video and announced on facebook that a new herd of the Auerrind project has been assembled in the Wistinghauser Senne in North-Rhine Westphalia. 

The video shows three cows and one young bull. I suspect that the bull is one of the two (Sayaguesa x Chianina) x (Sayaguesa x Watussi) bulls EDIT: It is the (Sayaguesa x Grey) x (Sayaguesa x Watussi) and the cows are most certainly Maremmana x Sayaguesa EDIT: Plus one Tauros cattle cows. That means the offspring will be both parentally and maternally a cross, which is good because that is where the real breeding starts. The offspring will have much Sayaguesa influence, which is good as well, plus, with luck, genes for large and long horns (Watussi and Maremmana), maybe the little bit Chianina will also help to achieve a large size. The young bull resembles his father Alvarez (Sayaguesa x Watussi) in the video. I am looking forward to see what his horns are going to be like as much as his final body size. 

Reconstruction model of the Indian aurochs

After I finished my reconstruction of the Sassenberg bull, I now can present another model, this time of the Indian aurochs. 

In the lack of a complete skeleton or an assessment of the known material, this necessarily involved more speculations than the reconstruction of the well-known European subspecies. It is based on several incomplete crania of presumably male specimen, mainly on two skulls which can be seen here. They are not complete, but some morphological differences to the European form can be derived from it. More speculative was the postcranial body, where I simply took the skeleton of the Store-damme aurochs as the base. Not out of carelessness, but because I assume there were no noticeable differences in proportions and overall body shape between the European and the other aurochs subspecies because all of those traits were functional (the short trunk, the long legs, the presence of high spinal processes in the shoulder area forming a hump etc.). The horns are based entirely on the skull displayed in the Geological survey of India, because there are photos in frontal view and from the side of it, giving a good idea of how the horn cores look like in real. The colour is almost entirely speculative. What is likely is that the Indian aurochs had the E+ allele on the Extension locus, and that there was sexual dichromatism to at least some degree. This is the result:

 

For the colour, I used Deshi zebu bulls as a template. Those light areas between the legs are speculative, but not entirely baseless as it could also have been present in the African aurochs and even banteng have lightly coloured “armpits”. Should there ever be evidence against those light areas, I can easily paint over them on my model, but I doubt there ever will be. The horns look very large after adding the horn sheath to the cores, but the Indian aurochs had proportionally larger horns than the European subspecies. And some specimen quite possibly also larger horns in absolute dimensions. I gave my Indian aurochs a larger dewlap than what is likely for the European subspecies, because of thermoregulation and display function. Also the ear shape is that of zebus, because I consider it likely or at least possible that the ear shape of zebus was inherited from the Indian aurochs. This gives the model a more indicine appeal, additionally to the skull shape. All in all, I think it is a quite plausible representation for what the ancestor of zebus possibly might have looked like. 

 

Life reconstruction model of the Sassenberg bull

It has been two months since my latest post. I didn’t publish any posts in those two months because I was busy doing new aurochs models. I started models of the Sassenberg bull, an Indian aurochs, and the Gramsbergen skull fragment. And last week I finished my model of the Sassenberg bull. 

 

As usual, I started the model by working directly with photos of the original bones/skeletal mounts to achieve maximum accuracy. I know photos of the Sassenberg bull skeleton in lateral view, but the postcranial skeleton of this specimen is partly a composite, and thus not the most reliable options when we have more complete skeletons of a single individual. Therefore, I chose the Store-damme bull skeleton as the basis for the postcranial skeleton. I used a photo from the internet which shows the skeleton in a perfectly lateral view, and “corrected” the inaccurate posture of the mount using GIMP. I manipulated the position of the skeletal elements until an accurately assembled skeleton was the result. Using a different bull for the postcranial body is not a big issue, as there was little to no variation in the postcranial proportions of the aurochs, except for the fact that some had longer spinal processes in the shoulder region than others, but not to a huge extent. For the anatomy of the skull/head, I used a photo of the Sassenberg bull. The most idiosyncratic part of an aurochs were its horns, and I used photos from several different angles in order to correctly replicate the horn cores. Then, after sculpting the body, I sculpted the horn sheath on the horn core. This is necessarily only an approximation, as there was no general rule how much the horn sheaths added to the length and thickness of the horn. For the body morphology, I did not use more or less strongly domesticated cattle breeds as an analogue, but chose Spanish fighting cattle individuals as much as living wild bovines. I did so because the morphology of cattle was greatly affected by domestication, and Lidia cattle are probably the least-derived taurine cattle breed. But there is always some variation concerning the bulkiness of the individual, and it also depends on the individual age. A very old bull, be it Lidia or a wild bovine, is certainly more massive than a very young adult bull. I wanted to reconstruct the Sassenberg bull in its “prime time”, perhaps at the age of around 6 to 7 years in domestic cattle age years. I will do more massive “old” bulls in the future as well. 

This is the result: 

The model was sculpted from air-drying modelling clay of the DAS brand and was painted with acrylics. It is 33 cm tall at the withers, the scale to an actual aurochs bull depends on which height you choose for that, as the European aurochs varied from 160 to 200 and perhaps more cm at the withers in life. During sculpting my model, I constantly checked its accuracy by superimposing a photo of the model on the “corrected” photo of the actual skeletal mount, and I can happily say that the model lines up perfectly with the skeleton in lateral view. As for the head and horns, here is a photo of the model next to a photo of the Sassenberg mount in exactly the same position: 

I am very happy with the result and I think it is quite likely that the model is more or less accurate. More reconstruction models are about to follow, the next one will probably be the Indian aurochs.