Created by Karl Pearson
1913
Description:

‘In a recent German Dictionary of the Natural Sciences there is an article by Dr Eugen Fischer on Race and Race Formation [E. Fischer, ‘Rassen und Rassenbildung, Rassenmorphologie, Rassenpathologie, Rassenphysiologie’, in Gottlob Eduard Link et. al. (eds.) Handwoerterbuch der Naturwissenschaften (Jena: G. Fischer, 1913) https://catalog.hathitrust.org/Record/012312933]. Dr Fischer is not the hack writer one so often finds as a contributor to encyclopaedias and dictionaries; he is an anthropologist of distinction and the opinion given by him represents fairly the current scientific opinion in Germany, and I fear in England also as far [as] we possess in such matters any independent opinion, - on the origin and decline of races. I venture to translate for you Dr Fischer’s section on Final Result of Race-Crossings: “A specially important question for the understanding and judgement of so many modern anthropological groups is that concerning the final result produced by race crosses. Does a new race arise? A mixed race with blended or new characters? Or, does one race more or less dominate? This last question has – so far as it concerns heredity – been already answered above. (The answer Dr Fischer has previously given is a reference to the papers of Bean, Salaman and Davenport, on dominance in [1-2] hair, eye, & facial characters.) Again the answer to the first question has been made by thousands of hybridisation experiments of the botanists and zoologists – no new race is to be expected. Characters always separate out again according to the Mendelian rules, the unit-characters are always found in thousandfold combinations alongside each other. The total population remains equal to the so-called F2 –generation (the second generation of hybrids) it is even to the percentage frequency of each character identical with the grandchildren of the pure race-ancestors. Of course it is supposed that no selection of special characters takes place and that there is completely random mating. In the case of man we are absolutely certain that it cannot be otherwise... The characters preserve themselves, an immense number of combinations arises, but no new race... Von Luschen was the first to emphasise this process, he emphasised how the old types always reappear, types which existed in a country thousands of years ago; Race mixture leads he said to a ‘sorting out’ (Rassenmischung zu einer ”Entwischung”). To-day when we know Mendelian segregation this is intelligible without any further explanation.[“]

So far Dr Fischer. According to this view man for thousands of years has consisted of the same component unit characters, and all that has taken place is endless shuffling of the same series of unit characters kaleidoscope fashion. The individuals who carry any given type of character may be more numerous now than in the past, but that type existed from all time, or if it has appeared at any stage in history, it has appeared as a sudden sport or mutation, which seems to be little more than a new name for the old ‘special creation’ of each type. Either man as we know him existed longside [sic] Palaeanthropus [sic], or he is compounded of unit characters which were all contained in [illeg.] Neanderthal man, and have been merely reshuffled, or he is the product of a “mutating fit” which has occurred since. Whichever view you take the whole conception of Darwinism the gradual evolution of man by the selection of small variations suitable to his environment is swept away. Neither theologians nor metaphysicians have been such opponents of Darwin as the Mendelians, and this [3-4] because the latter have largely started from the Darwinian camp & professed to be of its kind. The day is not far distant when the alternatives before us will be clearly recognised, we have to reject Mendelism or give up any consistent form of Darwinism, and that opposition between Mendelism & Darwinism is well known to the leading Mendelians even if they refrain from avowing it. The time, they probably consider, is not even yet ripe for such an avowal. Meanwhile memoir after memoir is published professing to demonstrate that all sorts of characters from plants to the lower animals and from these to man “Mendelise.” These memoirs are very largely of the type criticised by Dr Heron in his recent lecture on mental defectiveness – plastic material is made to fit an elastic theory in the manner of Dr Davenport. When it is convenient, a whole group of individual characters – each unmeasured - is treated as a whole, thus we have Salaman’s conception of the whole face treated as Jewish or Gentile, and we are told that Jewishness is recessive to Gentilism. When a character on the other hand does not Mendelise, we are told that [4-5] it is because we are dealing with a whole group of Mendelian units & that until they are separated out, it is not possible to determine whether Medelism applies or not. Not so many years ago we were informed that albinism in man was a Mendelian character. Dr Davenport even maintains it at the present day. Professor Bateson discretely [sic] remarked in the discussion on Heredity at the Royal Society of Medicine in 1908, that he had shewn pedigrees of albinism as an example of a character which did not follow Mendelian rules, but that those [note: ‘Proc. R. Soc. of Medicine Vol. II, 1909, p. 58.’] rules did apply to albinism in many animals. Personally I have shewn you in my second lecture when dealing with material at least twenty times as large as Dr Davenport’s, I can find no confirmation of Mendelism in human albinism. Another authority ships out all difficulty in the following manner:

Among albino rats, for instance the author of this article has reasons to believe, upon theoretical grounds resting on an experimental basis, that probably no less than thirteen types exist. With rabbits and mice there must be a still larger number. [note: ‘Encyclopaedia Britannica Art. Albinism p. 510, Vol. I, 11th Edn.’]

Perhaps in dogs accordingly there may be 50 or 100 a type, for every possible case of offspring that a posteriori are found to occur! If that is a philosophical account of albinism, I personally should prefer the facts and an absence of all theory whatsoever! But the [5-6] matter is much more serious than may appear on the surface. The grave fact that man can mould the future of his own race by marriage selection and rejection is becoming more & more recognised, and the patriot and the status-man of the future will turn to science in order to ascertain, where offspring are desirable and where not. Are we, or are we not, to pledge ourselves to the fashionable view propounded by Dr Fischer that race mixture leads to nothing but reshuffling of characters, which has existed in any country for thousands & thousands of years, and that any novelty, if indeed there be such, is due to a sport? Well, I think, before we accept that view, we should not only criticise the sort of manner in which writers like Davenport reach Mendelian results and apply them to the most important social problems, but we should try and find out for ourselves by examining new data whether Mendelian rules do apply to the lower animals and can then be transferred to human beings. The topic of my lecture tonight is going to be hybridisation in dogs. It is not my purpose to show you pretty animals & ask you to admire them. I want you to [6-7] ask yourselves are these data consistent with Mendelism, and if not, what safety is there in applying Mendelian rules to human beings, and telling the sane offspring of the insane to marry into normal stocks?

In my second lecture I remarked to you that in association with my two colleagues Mr E. Nettleship and Mr C.H. Usher, both well known ophthalmological surgeons, I started an inquiry into human albinism. That inquiry developed almost beyond our powers of coping with it, and one of the most important sides of it was the comparison of albinism in animals with albinism in man. Owing to a lucky chance Mr Nettleship was able to purchase the albino bitch Tong I, who eventually passed to me; she was followed by his purchase of two albino puppies Jack & Jill. Later Spook was obtained who passed to Mr Usher, and he in turn purchased the albino bitch Beenie. These dogs are all pure bred Pekingese of the best strains sired by prize dogs & Tong I, Jack & Jill obtained prizes themselves. From these fine dogs, our foundation stock, we have in [7-8] the course of 4 or 5 years had 56 albino dogs besides crosses with normal dogs leading to upwards of 50 non-albino dogs. Photographs of these dogs have been taken; on death their skins have been preserved, and all the chief internal organs for typical cases examined; the eyes have been carefully sectioned and the hairs microscopically examined. For the facts and material I show you tonight I have largely to thank my colleagues. You may take it that they are quite unbiased; indeed Mr Nettleship is, I take it, a Mendelian, and I do not know whether Mr Usher is or is not another. I, a non-Mendelian do not distrust any of the facts upon that account & I think you can trust them also. But I release my colleagues wholly for [sic] any responsibility for my conclusions. It is possible that they would not accept them or would differently interpret them. For myself, I feel convinced that these facts will not bear a Mendelian interpretation and I shall ask you to examine them from that standpoint.

In the first place, as far as Mr Nettleship has been able to ascertain, albinism first appeared in Pekingese in the offspring of [8-9] champion Goodwood Lo, a red dog with black mask, from Ah-lu-ta, a Golden sable. Goodwood Lo himself is descended from Goodwood Meh and Ah Cum, the former was mahogany red and the latter has been described as “bright sable red”. He was imported directly from the Palace at Pekin in 1896 by Mr & Mrs Douglas Murray and his skin is set up in South Kensington museum as that of a typical representative of the Pekinese [sic] breed. Our dog Tong I was a granddaughter of Goodwood Lo & Spook the whitest of all our albinos was Tong’s granddaughter. The second pair of our foundation stock, Mr Nettleship’s Jack & Jill were descended from Goodwood Lo & from Ah Cum & Meh. Beenie the last acquisition of Mr Usher was a granddaughter of a normal brother of Jack’s. The actual coat colour of these dogs varies from a light yellowish cream to pure white, much whiter in the Tong group than in the Jack & Jill group. This yellow tinge is very common with human albinos; it does not arise from pigment granules in the hair, but from a diffuse, probably lipochrome colouring substance. In the bulk of these dogs there is no melanin pigment in the coat at all; [9-10] in others, but only so far in the Jack & Jill group there are “foreign bodies” here & there, and it is a disputable point whether they are accompanied by granular pigment or not; in still others an occasional granule can be found; but there is nothing beyond what can be found in the hair of albinos of various native human races. In the internal organs, including the heart, lungs, hair, spleen, brain, internal ear, etc where pigment is found in the normal dog no pigment was found in the mesoblastic tissues of iris, ciliary body and choroid; the retinal epiblastic of iris was lightly pigmented, there was also pigment in certain other parts of the eye. The eye contains about the same amount of pigment as we have found in the eye of an albino negro, and it is apparently the only pigmented part of the animal. But one point is quite clear, i.e. that as in the case of man, albinism is not a unit character, but actually its amount differs from dog to dog & various grades of albinism exist. Even the coats [10-11] of some faun & red Pekinese with dark eyes which pass for normal dogs have no pigment granules ion their hairs, while Mr George Brown’s cream dogs with black muzzle & dark eyes – Huang-tu & Pailu are perfect albinos of the coat.

The causes which have produced albinos in the Pekinese & different grades of albinism are unknown to us, but it is easy to find such cases arising in the wild dog, the Dingo of Australia. I owe one such dog to Mr Tyrwhitt[-]Drake, whose pack of Dingos has shown recently incomplete albinism, a second case was sent to me direct from Australia with a normal skin & the eyes of the albinistic & normal dogs. If we place the skins in order of presence of red pigment, we find that this is the order of the amount of pigment found by Mr Coates in the eyes, the albinism of the eyes increases with the albinism of the skin. We have albinism spontaneously appearing in the wild dog & the colour of the coats approaching nearly to the cream of some of our albino Pekinese. The important point here is to mention that there is no rigid line between albino & normal, it is a quantitative step. [11-12] That is precisely what I have found in the case of man.

Now the dog has many characters which we can study from the standpoint of heredity. We can in this case take the degree of albinism of the eyes, the colour of the coat, the length of the hair, the nature of the muzzle, or the shape of the head, besides a variety of psychological characters, which everybody who studies dogs will at once recognise. In our work with dogs we have crossed the white coated albino Pekingese with the black coated Pomeranian. No greater contrast within the limits of easy matings could be imagined. The Pekinese is a royal dog with lordly habits; he is fearless, self-willed, keen on scent, and may even be hunted in a pack, he is intelligent & of cleanly habits. Compared with him the Pomeranian is a cockney upstart; he snarls rather than barks and is almost invariably a coward, of doubtful habits. He will admit defeat at the claws of a cat or consent to share a hearthrug with her, which none of my Pekinese would ever submit to, indeed I have seen them roll over and put to flight a collie dog who had behaved disrespectfully, although they will not seek a fight. [12-13] If you tread on the tail of Ling the dog in the cage there, he will not howl, but growl & retaliate if he can by taking a piece of your toe, even if you are his master. Later he will apologise handsomely. A Pomeranian would have howled & then fawned upon you for comfort. It is singular to see in the hybrids some of the psychical characters of the Pomeranian appearing under a physically Pekinese exterior.

I propose to consider the several hereditary characters of these dogs.

First as to the albinism. Does it obey the laws of Mendel? In the first place we must consider, whether an albino mated with a coloured Pekinese gives all albino. Here is such a mating, VI.24 x VI.46, all the offspring are coloured dogs. But all the offspring were coloured dogs. But all the offspring need not be coloured; here is such a mating VI.31 x VI.38, the offspring were these albinos and two coloured dogs. The first result tells us that albinism in Mendelian language must be a recessive character, the second result the Mendelian explains by asserting that VI.38 carried latent albinism, while he would suppose VI.46 did not. He cannot contradict this, because V.38 was a red dog called Ping Pong and we know nothing of its ancestry. VI.46 was a black dog imported from China in 1907, and the history of neither can be followed. Here arises our first difficulty. In a stock like this all going back to the Palace in Pekin it is impossible to assert that a dog has or has not albinism latent, i.e. it is a (DR), we simply call him a hybrid as far as albinism is concerned, when he has any albino offspring and when he doesn’t have any we at once call him a dominant normal a (DD), but change him to a (DR), if an albino turns up later. In order to surmount this difficulty we have introduced the pure black Pomeranian. I never heard of an albino Pom, and we may be certain that our Pomeranian dogs are (DD)s or when crossed with the albino have all their offspring (DR)s. We have therefore to consider the possible matings of (DD), (DR) & (RR), namely:

(i) (DD) x (DD) = 4(DD) – i.e. all normals

(ii) (DD) x (DR) = 2(DD) + 2(DR) = 50% normals + 50% latents

(iii) (RR) x (RR) = 4(RR) i.e. all albinos

(iv) (DR) x (DR) = (DD) + 2(DR) + (RR) = 25% albinos

(v) (DD) x (RR) = 4(DR) = all apparent normals

(vi) (DR) x (RR) = 2(DR) + 2(RR) = 50% albinos.

[14-15] It is impossible to test (i), because whenever a normal dog mated with a normal dog gives an albino we say both parents were (DR) or carried latent albinism.

It is generally impossible to test (ii) because if an albino did appear we should say that the (DD) was really a (DR) & transfer to (iv). It is the like with the (v) type of mating for if an albino did appear it would in its turn be transferred to (vi). Thus (i), (ii) and (iv), when we are not certain of our (DD) will tell us nothing for or against Mendelism.

Now let us look at the remaining three cases, where we actually are certain of our characters. First albino & albino. We have 16 litters of this type and around 66 puppies have been born. The anomalies that occur have really concern[ed] the extent of the albinism, and they occur in particular in the Jack & Jill offspring. Now you have seen what Jack & Jill are like. Here is the coat of a normal albino, practically an adult dog. Here again the coat of an albino puppy. Now what do you call this coat? It is very hard to assert that it is that of an albino dog, when its own sister at [15-16] the same age has a coat of this kind. Again here is another son Fo of Jack & Jill – he is practically a piebald, white & faun. Still more is this the case with Fi as you may judge from her photographs. Yet in order to establish Mendelism you must assert all these dogs to be albinos. If you judge by the eye, and not by the coat none of these dogs are complete albinos, just as very probably few human beings are complete albinos. Thus your dogs are producing various grades of albinism – less intense than their own, just as the normal dingo in its turn produces variations tending to incomplete albinism, from which in all probability more marked grades of albinism could be reproduced. Those who to fit a theory in their obsession cram all these dogs into one category & speak of albinism as a unit character, are doing as much harm to the real study of heredity, than those who speak of mental defect as a Mendelian unit character, while the[y] screen[?] under it epilepsy, neuralgia & god knows what not.

I now turn to the mating of (DD) with (RR) of normal dog with albino dog. As I have said this mating must give no albino, because when it does the normal is at once stated to be a “latent albino.” The only case where we can be sure of our clarification is when we mate the albino Pekingese with a black Pomeranian. Three such Pomeranians have been used Olga, Dido & Prince. Mr Usher chose these dogs because they have as black an ancestry as it is possible to get in the case of the Pomeranian. Olga has both parents black, all four grandparents, all eight great grandparents, all 16 great, great grandparents, and of its 32 great great great grandparents, 28 were black, 2 were brown and the colour of 2 unknown. Prince is the son of Olga & a black dog Janos[?] and all his ascendants up & including his 32 great great great grandparents are all black & it is not till we get to the sixth ancestor in ascent that we touch a brown dog. Dido is a less pure bred black. She has one brown grandsire and one brown great great granddam. From Dido & Olga Mr Usher has had 2 and 4 litters respectively including twelve puppies from albino male Pekinese. From the albino bitch Beenie and Prince there has so far been one litter with four puppies. Now what [17-18] might we expect? Albinism is recessive, therefore we should expect to get black dogs with dark eyes. The hybrid we call the Pompek. It is a black dog with dark eyes, but in almost every case there is a white shirt front. Where does that shirt front come from? It is hardly deduced from the Pom ancestry, which has had whole colour dogs. There is in my mind little doubt that the alternative ancestry constitutes the white shirt front to the Pompek. Donald dhu the Pompek I have here tonight is nearly a whole black dog, but my photographs will show this white shirt front. Are you going to pass that shirt front wholly by and talk about dominance of the black? But this is not the end of the matter - besides these black & white Pompeks there have appeared in the litter by Prince from Beenie not brown dogs, but two chocolate dogs, & yet Prince has not even a brown dog in his ancestry until you get to the 6th ancestral generation. We are informed that pure bred black Poms crossed with pure bred white Poms will give chocolate dogs. In other words, even if we pass over our white shirt fronts, we find the possibility of a blend in the chocolate Pompek. It is [18-19] difficult to understand what is meant by the statement that albinism is recessive in such cases. It is true that no perfect albino has arisen in the first generation, but it is equally true that no perfectly black dog has arisen, for even in Donald dhu it is possible to find some white hairs, and he is exceptionally black. I have no doubt ardent Mendelians would tell you that this is a case of imperfect dominance. But what is the good of a rule of dominance at all, if a careful inspection of your hybrids leaves you in doubt as to how to classify them? If the offspring of a sane individual and a mentally defective woman is markedly neurotic or hysteric, what purpose is gained by talking of mental defect as a recessive character?

Next let us turn to the mating of latent albino with latent albino, i.e. of (DR) with (DR). These matings should give one quarter albinos. I have already pointed out the difficulty of settling among these Pekinese what is an individual with latent albinism. If we take only the test that two dogs shall have produced one albino we are excluding those cases in which two (DR)s produce no albinos. In particular [19-20] our pedigree as a rule only contains only those litters of two given dogs that do contain albinos. Of course the Pompeks are all certain (DR)’s. Putting all the available evidence for Pekinese together we have 11[19?] litters producing 46 puppies of which 14 were albinos – not a very good Mendelian quarter but not so bad that any argument against Mendelism could be based on it. At present we have only succeeded in getting two Pompek & Pompek crosses, leading to 8 dogs. Of these eight dogs six are black with white shirt fronts, thus resembling with perhaps a little more white the first cross, one is a dog, Sheila, with normal eyes, but a sable coat; in coat she resembles not the albino, but the normal Pekinese and it is quite impossible to say why she should come in at all on the Mendelian system; and the eighth dog had albinotic eyes, but a cream coloured coat with a white shirt front. If the appearance of a dog with black muzzle sable coat & normal eyes together with an incomplete albino be evidence of an albinotic character, then the experiment provides the required Mendelian result, but personally I wholly fail to grasp in what Mendelian category Sheila is to be placed. She is not an albino or [20-21] recessive; she is not a dominant for that is a whole black coated Pom, and she is not a hybrid or (DR) for that is a black dog with a white shirt front. She is a reversion to a fairly familiar type of Pekinese, who has no business in this Mendelian scheme at all.

Lastly let us consider the mating of albino (RR) with latent albino (DR). We have at present only 5 such matings giving 7 albinos and 17 normal dogs – a very poor ratio of equality. But the matter is much worse if we confine our attention to the cases of certain (DR) namely the Pompek with Pekinese albinos, we have bred at present 15 such dogs. Only 3 are albinos of a kind, and 12 are non-albinos – a remarkably poor half! Of these fifteen dogs six are of the (DR) type i.e. a black dog with a white shirt front; a seventh possesses not only a white shirt front but one white paw, two only are complete albinos; two are red dogs reverting to the normal Pekinese; two are perfect piebalds, i.e. black & white in patches as you will see by the skins, and one is a lilac skewbald with red eyes. The appearance of the piebalds when the [21-22] hybrid is mated back with one of the original stocks we have already referred to in the case of man. There are no piebalds, as far as we are aware in Pomeranians, they exist but are rare in the case of Pekinese & although the tendency to piebaldism appeared in the offspring of the albinos Jack & Jill, we have no record of it in our Pekinese normal ancestry. The results with their reversions to distant ancestry seem far too complex for any Mendelian simple formula of dominant & recessive characters to cover them. They correspond exactly to the wonderful breaking down of pigmentation which occurs in the case of man, when a dark race is crossed with an albino and the offspring again mated with one of the parent races; not only albinos & blacks may appear but piebald negros, & xanthous negros with yellow or red hair. Albinism is not a simple Mendelian unit character it is a quantitatively graded character, & from albino & melanotic individuals in a race almost every single shade of colour & every variety of piebaldism can be produced by judicious crossing.

So much for the colour results of these [22-23] dog breeding experiments, which of course are still far from complete.

I now turn to the shape of the head and the muzzle. There can hardly be a greater contrast than [the] broad head of the Pekinese with its short muzzle and the narrow head & long muzzle of the Pom. One felt at once here was a very marked character which in its different forms was hereditary in the two species, and which ought to be a very good feature to test Mendelism from. The first cross between Pomeranian & Pekinese were all described by us as “short muzzled dogs”, and I must confess that I spoke very glibly myself of the “dominance” of the short muzzle. The long fine muzzle of the Pom was clearly recessive, and it ought to come out again when the Pompek was crossed with Pompek. It cannot be said to have done so; what we have got are rather mongrel looking heads which are neither Pom not Pekinese. It then dawn[ed] on me that these things are not to be described by such vague Mendelian categories as short & long muzzle & that it was absolutely needful to measure all the dogs’ heads before one could be certain [23-24] what was taking place. Accordingly I arranged a rough scheme of measurement for the dogs heads. It must be rough, for we take it on the living dog, and it is often not an easy task to apply the callipers to the head of an active & self-willed young dog. Still I am quite sure these measurements are far more valid than any mere appreciation of long or short muzzles. I call for our present purposes the total length of the head the distance from occipital ridge to the point in the sagital plane of the middle line of the nostril, from the lather[?] point to the bridge of the nose, as near the [N]asion as I can get on the living dog, I call the nasal length. The maximum breadth of the head is found on the zygomatic arches; and lastly the breadth of the jaw is taken just under the eyes on the upper jaw. I then make these indices;

The head index = maximum breadth/maximum length

The muzzle index = length of nose/breadth of jaw

The nose index = length of nose/maximum length of head

It is the heredity of these three Indices that I have investigated & we inquire whether they “Mendelise”. If I am told that they are very [24-25] complex characters I reply that they are simplicity itself compared to such a vague entity as the “Jewish race,” or indeed to a more general application of Pekinese or Pomeranian muzzle – or again to piebald or non-piebald

We have the following mean results:

Head index Muzzle index Nose index

Pekinese 81 48 27

Pomeranian 68 102 42

Thus not only are the means widely different but the ranges in the dogs of both races we have measured in no case overlap. Now what happens when we cross two Mendelian characters? According to Mendel one will dominate.

Now here are the head indices of 9 Pompeks. I should be glad to know whether the long or short head has here “dominated”? The highest head index in a Pompek is the lowest we have observed in a Pekinese & the lowest head index in a Pompek is the lowest we have yet observed in a Pom. In other words, if segregation means the occurrence of characters in the original stocks, it occurs already in the first generation of hybrids. There is no possibility of speaking of dominance.

When the Pompek is crossed with Pompek we find the range of variation still further extended, it goes from 58 to 83, instead of from 65 to 79. But if we cut off the Mendelian quarters at either end and say there are Poms & there are Pekinese, we should be equally justified in asserting that the quarters in the first hybrid generation are Poms & Pekinese, for they fall practically outside the same divisions. The whole process would be simply arbitrary & made in order to reach Mendelian quarters & a great deal of the present Mendelian dichotomies are precisely of this character.

Now take the muzzle index. Here we have in the first generation of hybrids one muzzle index which stands right below the Pekinese and a range from 41 to 77. Are we to call this “dominance” of the Pekinese muzzle? Why 8 out of the 9 Pompek indices are in excess of anything we have found for Pekinese, which range any [sic] from 45 to 50! We might just as well fix 50 as our limit & term it “dominance” of the long muzzle! Now we come to the result of mating Pompek with Pompek, 7 of our second generation of hybrids now lie gain within the Pompek range & one is pulled up without quite [26-27] reaching the pure form range – this is Sheila. We are not justified in asserting that the first two are Pekinese & the last a Pom: for on the same evidence the dog with a muzzle index of 41 in the first generation must be a Pekinese & segregation have taken place in the first generation of hybrids.

Precisely the same result flows from our returns from Nose Index. The three largest Pompeks are already Pekinese, and the highest Pompek a Pom. In the second generation of hybrids, two dogs instead of three fall into the Pekinese range, but two dogs instead of one fall into the Pom range. There is thus little more segregation in the second generation than in the first. Only further experiments will show how far any of these extremes will breed true & there are not at present enough extremes alive to test this point.

In the next series of crosses we have made that of albino Pekinese with Pompek, the same difficulty of interpretation arises. If the Pekinese character were dominant, all the offspring should be like the Pekinese; 50% will breed true, 50% will not. If the Pekinese character be recessive then 50% should be like the Pekinese [27-28] 50% like the Pom. This is true of any single one of the characters.

If we take the muzzle index – none are like the Pom, therefore the Pekinese muzzle must be “dominant.” But only two fall below & are just above the Pekinese range. 30% at most is a poor 50%.

In the case of the head index, two dogs exceed the Pekinese range, but the crosses cover the entire range of Pekinese, Poms & Pompeks. It is impossible to say that 50% are Pekinese & 50% Pompeks, nor can we assert dominance of either Pekinese or Pom character.

Absolutely the same applies to the nose-index the results obtained cover the whole range of Pekinese, Poms & Pompeks. We can assert no dominance and no 50% of either Pekinese or Pomeranian element.

In fact the moment we begin to measure characters in the dogs and not to take vague appreciation of Pekinese or Pomeranian characters we fail to find any approach to Mendelian rules.

What rules do then approximately hold for these measurable characters? Why they simply seem to blend in the manner that Sir Francis [28-29] Galton showed us years ago holds in the case of measurable characters in man. The accompanying slide brings this out markedly. The one anomalous results is in the muzzle index of the actual Pompek. This is intermediate between the muzzle indices of Pekinese & Pomeranian, but is not half-way between,- 61 instead of 75. It is conceivable that a larger number of Pomeranians measured will show that 102 as the mean muzzle index of Pomeranians is too high.

Mendelism has become the mode – no other conception of heredity can today obtain a hearing. Yet, I believe, the time will shortly come when a reaction will set in, and the views of Galton will again come by their own. At any rate the present experiments on dogs seem to me to indicate that there is still a chance for philosophic Darwinism. Even by hybridization, a new race can be created which is not a mere shuffle of old unit characters, but is a true intermediate. It remains to be seen whether new forms lying outside the original Pekinese & Pomeranian ranges cannot be perpetuated – say a black dog with a white shirt front a head of over 72, a muzzle index [29-30] of 60 and nose index of 32. Such a dog possesses characters which do not appear in the parent breeds. It is far too early yet to assume with Dr Fischer, that nothing new can come into the world of mankind, because botanists and zoologists have answered by “thousands of hybridization experiments that no new race is to be expected, but only a shuffling of old unit characters. “ Thus far these hybridization experiments have not been conducted solely with a view to finding out the facts, they have been dominated by a theory, to which everything else is recessive. I believe when the callipers receive general application, and the record of facts is undertaken by those without theory then there will be some possibility of testing whether the Darwinian philosophy is to dis-appear before a theory which provides nothing but a shuffling of old unit characters varied by the rare appearance of an unexplained “fit of mutation.”

Meanwhile I venture to assert that at least dogs can demonstrate for us that the application of Mendelism to anthropology and to social problems is wholly premature and [30-31] that rules as to disease and pathological states, which have grave social bearing as for example that mental defect is a Mendelian recessive, or that normal numbers of insane stocks may safely marry into healthy stocks – ought not to be propounded as guides to human conduct. If there is to be a real science of Eugenics then we must at least avoid spanning the crevasses in our knowledge by a snow bridge of theory. A record of facts lasts for all time, but theory is ever in the making or unmaking, - a mere fashion which describes more or less effectually our experience – to extrapolate it beyond that experience in nine cases out of ten leads to failure – to disaster when it touches social problems. In all that relates to the evolution of man & to the eugenic problems of race betterment, it is wiser to admit our present limitations than propound like Drs Fischer & Davenport sweeping Mendelian racial theories. Let us as Eugenicists adopt the tone of the soothsayer in Anthony & Cleopatra, and when we are asked “Isn’t you, sir, that know things?” reply: “In Nature’s infinite book of secrecy, a little can be read.”’