Chapter XIII: Part III: Family Studies in Heredity of Build (6)

1. Male (III 3), a clerk, has at 66 years the formula 150/69 inches; build 2.2 (32).

2. Male (III 4), a physician, who has always had a rheumatic tendency and is liable to sciatic neuritis, at 25 years had the formula 180/68.75 inches; build 2.7 (38). He is the Father of Thr-2 family, page 79.

3. Male (III 6), a preacher, has at 69 years the formula 170/68 inches; build 2.6 (37). He has a daughter with a build of about 3.2 (45).

4. Male (III 7), a physician, who in middle life suffered from inflammatory rheumatism, has at 59 years the formula 170/68 inches; build 2.6 (37).

5. Female (III 8), who died at 34 years from appendicitis, has the formula 225/67 inches; build 3.5 (50). She was “a big overgrown woman in her teens.”

6. Female (III 9), married, who has a rheumatic tendency, has at 55 years the formula 190/64 inches; build 3.3 (46).

7. Female (III 11), married, who has 9 children, suffered in middle age from rheumatism. At 53 years her formula was 225/67 inches; build 3.3 (46).

8. Male, a merchant, well, has at 51 years the formula 155/67 inches; build 2.4 (35).

9. Male, a merchant, well, has at 49 years the formula 165/68.5 inches; build 2.5 (35).

10. Male (III 14), a doctor and dentist, has at 47 years formula 180/68 inches; build 2.7 (39). One of his daughters has a build of about 3.2 (45).

Child No. 2, a physician, writes: “Many of my relatives, both on Father’s and Mother’s side, were afflicted with so-called muscular rheumatism (cousins, uncles, and aunts), pains in the limbs and muscles (not in the joints); especially those of robust and fleshy build, who possessed extraordinary good appetites and digestion, who liked good things to eat and much of it; the women being good cooks and the men generally marrying good cooks, they always had plenty to eat (rich food) and to spare. None of them are users of alcohol, simply being extra-hearty eaters” (R: Thr-1).

WEN FAMILY. (Fig. 49.)

IV 10 is the Father; his formula is 260/74 inches; build 3.2 (46); very fleshy. His father (III 10) had the formula 245/73 inches; build 3.2 (46); very fleshy. This father had a brother of just his build (who eventually died of tuberculosis) and a brother of formula 250/74, build 3.0 (43); fleshy. The Father’s father’s father (II 5) had the formula 200/70, build 2.9 (41); fleshy; one of his brothers was fleshy (38), and one very fleshy, build 3.2 (45). The latter had a very fleshy son. The Father’s father’s father’s father had a build of 2.9 (41) and a brother of the same build. These brothers of Gen. I were products of a first-cousin marriage. The Father’s father’s mother was a woman whose formula was 200/68.5 inches; build 3.0 (43); fleshy. Hence the Father’s father and his brother, both of build 3.2 (46), got the fleshy tendency from both sides of the house. The Father’s mother was a second cousin of the Father’s father and her line also contained many fleshy representatives, but she herself was of medium build. Thus the Father belongs to a fleshy strain, but he probably carries some non-fleshy gametes.

IV 11 is the Mother; her formula is 160/65.5 inches; build 2.6 (37); fleshy; she has 2 sibs, both fleshy; her parents are of medium build; they have some slender grandchildren.

There are 8 children, V 3 to 12, whose sex and indices of build are as follows: ♂ 3.6 (51), ♀ 3.1 (44), 2 ♂ ♂ 2.5 (35), 3 ♀ ♀ 2.3 (32), ♀ 2.2 (31). Thus 2 are very fleshy and 6 are of medium build. The interval between the two groups is a wide one, indicating a clear genetic difference between them. The very fleshy brother, who weighs 265 pounds, was able to reduce about 60 pounds to be accepted for army service, but since being discharged from the army he has rebounded to his former weight. (A:032-25.)

SUMMARY AND DISCUSSION.

In a simple monohybrid with complete dominance the back-cross upon the recessive gives 50 per cent of the heterozygous dominant and 50 per cent of the recessive allelomorph. In a monohybrid with imperfect dominance the back-cross upon the recessive gives 50 per cent of the intermediate and 50 per cent of the recessive. Where 2 or 3 multiple factors are present and the heterozygote is nearly intermediate but shows a slight tendency toward imperfect dominance, we may expect an approximation to a symmetrical distribution of grades about the intermediate condition, with a skewness toward the recessive condition. That is what is got in both parts of table 32. This result supports the conclusion that there is segregation in the gametes of the heterozygous parent, so that the progeny tend prevailingly to fall again into the slender and medium-fleshy grandparental types. The results of table 33 similarly support the conclusion of segregation in the gametes of the heterozygous fleshy parent.

C. GENERAL DISCUSSION.

HEREDITY AND ENVIRONMENT IN BUILD.

That a tendency to slenderness or fleshiness of build “runs in families” and characterizes different races is a matter of common observation. But this fact is far from satisfying the clinician that heredity plays any part in this result. Thus, von Noorden, who occupies a leading position among Teutonic investigators of metabolism in general and obesity in particular, denies the importance of an hereditary anomaly of metabolism in different families and races. Rejecting “anomaly of metabolism” or “peculiarity of protoplasmic metabolism,” he stresses “inheritance” of habits of life that favor obesity, the quantity and quality of food, and the ideals of bodily activity. For example, the Eskimo are fat because they eat blubber and huddle in narrow spaces, undergoing little movement throughout the long, dark winters. It is true that von Noorden speaks somewhat guardedly; he recognizes exceptions; suggests a possible hereditary hypofunction of the thyroid. But clearly constitutional peculiarities are, for him, exceptional as causes of overweight, and here is where he fails to recognize sufficiently the fact that usually only particular individuals of a fraternity are fleshy; the others may be slender.

But besides anomalies of protoplasm and family traditions of feeding, there are obviously other possibilities. It is well known that different varieties of cattle differ greatly in their capacity for fattening. Armsby and Fries (1911) have inquired into the influence of type upon the fattening of cattle. They used a pure-bred Aberdeen-Angus steer and a “scrub,” part Jersey, steer for comparison. As is well known, the former belongs to the easily fattening beef type; the Jersey to the difficultly fattening milk type. During over 2½ years, beginning at under 1 year of age, these steers were under nearly continuous observations. They were fed on ordinary growing rations, the same for each steer. The digestibility of the total ration was determined at intervals; four tests were made of each animal in the respiration calorimeter to determine the percentage availability of the energy of the feeds consumed by each. The results were as follows: Analysis of feces and urine failed to show any difference in percentage digestibility of the food by the two animals, and calorimeter tests failed to show any difference in the proportion of the food-energy which was being metabolized. But the two animals did not metabolize in the same way. Thus, in the scrub, a larger proportion of the gain made was of protein than in the case of the beef steer; and, conversely, the gain of the beef steer was more largely fat than in the case of the scrub. Reduced to common weight, the energy requirement for maintenance of the scrub steer was nearly 19 per cent greater than for the beef steer. Since the beef steer would eat more than the scrub and tended to store fat rather than protein, the greater tendency of the beef steer to fatten received a biochemical explanation. The results seem to show a difference between the two varieties in the working over of the assimilated materials.

Indeed, it is easily appreciated that steers of the beef and dairy types of cattle should metabolize differently when we consider the marked difference in the milk production of the cows of these two types. The cow of the highest dairy type is capable of manufacturing 20 kilograms of milk containing 1.2 kilograms of butter fat in one day, or 6 per cent (Bailey, Encl. Am. Agric., III, 365). The cow of the meat type, of larger size, produces up to 30 kilograms of milk, and this contains, perhaps, 1.3 kilograms of butter fat per day, or only 4.3 per cent (Sinclair, 1904, p. 740-42). There is here, obviously, a difference in the metabolic processes in the cows and this is reflected in the steers also.[11] There is an internal biochemical difference as well as a difference in the feeding instinct. The latter is not merely a matter of family tradition, of the family economics or mores; it is a physiological phenomenon as much as internal metabolism.

[11] I am indebted to Mrs. C. D. Walcott for calling my attention to this point.

Indeed, even von Noorden (1907, III, p. 700) seems to be forced to this conclusion by cases of failure to reduce weight at a diet far below that which appears to be essential to maintenance. Thus a man of 39 years, who exercised freely in the open air, had a weight of 102 kg. For 3 months his diet never exceeded 1,720 calories (estimated at 1,000 calories short of normal requirements) and at the end of this period he weighed 101 kg.

That a relation between ingested food and activity is not the entire explanation of obesity is recognized by medical men of experience. Heckel (1920, p. 371), referring to recurrent obesity, remarks on its frequency and says: To constitute an obesity there must be organic and hereditary tendencies: “Ne devient pas obèse qui veut.” “Aussi la guérison accidentelle ou thérapeutique d’une obésité n’indique-t-elle pas la disparition définitive des tendances personelles ou congénitale.”

Gulick (1922) has lately undertaken experiments to throw light on the question why some persons fatten easily and some with difficulty. He had noted that he himself belongs to a non-fattening strain and that his inclination toward a very copious diet of predominantly starchy nature did not lead him to put on weight, even though his round of activity was moderate. So he undertook biochemical, nutritional studies on himself. His observation covered nearly 21 months. During part of this time his caloric intake was low, 1,875 to 2,780; during another part high, 3,400 to 4,100. He found that he fattened somewhat during the period of heaviest feeding. There was, however, always an excess of intake over predictable need, and this excess increased absolutely and probably even relatively as the intake increased. The fecal nitrogen was 2½ to 3½ times greater during over-feeding than under-feeding. The basal metabolism during maximum feeding was normal. Gulick concludes that a person belonging to the difficultly fattening type shows a wasteful rate of oxidation, whether under or over fed, but especially at the latter time. “It seems clear,” concludes Gulick, “that throughout the entire experimental series there was some factor at work which caused fuel food to be burned more freely than in the average individual. This factor was not an over-active thyroid, as attested by the entirely normal basal metabolism.” Gulick concludes that it was “some factor in the chemistry of nutrition” which caused extravagance. This he thinks may very possibly be comparable to the “secondary effect” of protein enrichment, which, according to Rubner, can raise the specific dynamic action of the food without raising the basal rate. It is also possible, he says, that the spare type may be accounted for by any factor that produces a high “cost of digestion,” just as the obese may be supposed to suffer from an abnormally low “cost of digestion” (von Noorden).

Whatever the fundamental cause may be, the fact remains that in certain families there is a widespread inclination to the production of slender individuals, while in other fraternities certain proportions (though usually not all) of any fraternity are fleshy or even obese. Perhaps, as in the case of the Jersey as contrasted with the beef steer, the two kinds of individuals do not metabolize their food in the same way; some are spare and muscular, others lay on fat. In any case we can not disregard the constitutional factors in build.

Looking at the matter broadly, we can see that no other theory than that constitutional differences as well as nutritional differences determine build is sufficient to meet all the facts. In other species of animals we have precisely the same kind of differences between hereditary strains of slender and stout build that we have among humans. Thus, among dogs, the slender greyhound or Dachshund and the robust “Chow”; among horses the Thoroughbred and the stocky Percheron; among swine the “razor-back” and the Berkshire; among poultry the slender Leghorn and the stocky Cochin. Slender and stocky as racial traits appear quite as white and black do and they doubtless have similarly a genetic basis.

A careful study of the families described in this paper must convince anyone, it seems to me, of the importance of the genetic factors. The Fun. family is of a wholly different type from the Thr-1 family. Even if the latter eat more, it is because of a constitutional urge like the constitutional urge that leads the Aberdeen-Angus steer to eat more than the Jersey steer. Very probably the Fun. family metabolizes in a different way from the Thr-1 family, building more protein and less fat.

In other cases it is not the whole fraternity that is fleshy or slender. This is well illustrated by the Wen. family, where 2 of 8 of the main fraternity are very fleshy and all the others of medium build. We have here to do not merely with a family habit of eating, but a differential constitution that provides one-quarter of the children with a large appetite that leads them to eat heavily and manufacture fat and provides three-quarters of them with a small appetite that leads them to eat lightly and to manufacture protein instead of fat. Constitutional differences in the appetite and method of metabolism are the essential factors; and these are the things that are inherited. Only on such an hypothesis can we account for the clear evidence presented of constitutional factors in build—not always one only, but sometimes three or more acting together to produce the end result of obesity.

A scientific man, interested in nutrition, who has a build of 3.6 (51) himself and one of whose sisters has a build of 3.1 (44), whereas his 4 other sibs have builds of 2.5 to 2.1 (35 to 30), writes that his brother (of medium build, but about 75 inches tall) consumes daily about 2,700 calories; he himself and his fleshy sister about 2,500, and the others of his fraternity, who are of medium build, 2,000 to 1,800. He is a professional man, who does a good deal of office or laboratory work. He drinks about 3 to 4 quarts of water per day. His son, who is nearly 11 years of age, weighs just over 120 pounds, is 63 inches tall, and fairly fleshy. Though he is active, “never still a minute,” yet he has to be urged to eat, willingly misses a meal, and uniformly declines a second helping; “he eats much less than his 6-year-old cousins, who are actually under size for their age; he has never been a heavy eater from the time he was weaned.” This boy, the son of a very fleshy man, seems to afford an example of the easily fattening type in whom the slight excess of calories produces a striking result in build.

ENDOCRINE GLANDS AS LINKS BETWEEN CHROMOSOMES AND SOMATIC QUALITIES.

When it is concluded that there are constitutional, genetic factors for build, the validity of this conclusion is not weakened by the known rôle of endocrine glands on metabolism and build. Thus the experiments of recent years have emphasized the importance of the rôle that the thyroid gland plays in growth and development. As Uhlenhuth (1922, p. 182) says: “The ontogenetic development of the individual is controlled by the thyroid hormone.” Since in cretins the thyroid functions imperfectly or not at all, the differences between the development of a normal child and that of a cretin are supposed to indicate the scope of the control by the thyroid hormone. In the cretin, growth is slowed up almost to cessation. In the long bones the centers of ossification develop slowly. “The epiphyses may be absent many years after they are due to appear and their union with the shafts of the long bones indefinitely delayed. Periosteal as well as endochondral bone formation may be greatly reduced and delayed. The cranial bones are thin, poorly ossified, and osteoporotic: mottling is sometimes evident ontogenologically.”... Incomplete closure of the fontanelles is very common (Janney, 1922, pp. 391, 392). The genito-urinary system is undeveloped; the external genitalia remain infantile. The brain shows defective convolutions and all parts of the central nervous system show deficient development. Owing to the fact that the long bones of the leg develop imperfectly, while the chest may continue to enlarge, the build of the cretin is usually large for his age. Thus, in general, the specific metabolic changes are retarded. When the thyroid functions imperfectly after maturity the victim puts on fat. The fat is usually distributed nearly uniformly under all parts of the skin. One of Janney’s cases weighed 212 pounds at 12 years. Some degree of hypothyroidism is doubtless responsible for many of the “very fleshy” persons referred to in the preceding pages. The frequent recurrence of obesity in a family is in accordance with the tendency for hypothyroidism to recur in different members of a family; a tendency shown remarkably in Barrett’s (1919) family.

The hypophysis, especially its anterior lobe, is believed to exercise an important control over differential metabolism and the resulting form. The results of hyperfunctioning of the gland are very different from those of hyperfunctioning of the thyroid. An extraordinary enlargement of certain parts of the skeleton, especially of the base of the skull, the hands, the feet, and the ribs, occurs. There is, indeed, in growing persons a tendency to the production of giants. In the underfunctioning of the hypophysis before puberty an extraordinary obesity is apt to appear, known as dystrophic adiposo-genitalis; because, the more the fat increases in amount the more the external genitalia appear under-developed—infantile. The obesity frequently affects the region of the pelvis only or chiefly, though the thighs also may be involved. Disfunctioning of the pituitary in later years is believed, likewise, to induce obesity. Of this disfunctioning there are, doubtless, all degrees corresponding to the variety of grades of obesity. Thus a second factor of obesity must be recognized; and this also “runs in families”—is hereditary.

The other endocrine factors in obesity are less precisely known. It is believed that the interstitial cells of the gonads secrete hormones that regulate, especially inhibit, the production of fat. This is inferred by the frequent tendency of women after the menopause to grow fat and of eunuchs (of one type, at any rate) to become fleshy. Also, the secretions of the islands of Langerhans of the pancreas are known to influence carbohydrate metabolism, so that when they do not function properly sugar is eliminated unchanged in the urine, and “diabetes,” usually accompanied by fleshiness in its early stages, ensues. Of course, it can not be stated that pancreatic insufficiency is responsible for the prediabetic obesity; but the relation of the two phenomena is undeniably close.

Thus there come clearly to view two links in the chain of causes connecting chromosomal factors (such as are present in “Mendelian” heredity), on the one hand, and build, on the other. First, two or more of the endocrine glands play a great rôle in metabolism and when disturbed usually result in an abnormal build; and, secondly, just the condition of endocrine disfunctioning is an hereditary one and seems to depend on the condition of the enzymes or catalyzers that must be in the chromosomes.

That other constitutional conditions than those of the larger endocrine glands may play an important rôle in metabolism can not be denied; probably the quality of the protoplasm of every active cell influences the bodily metabolism; but the endocrine glands proper seem, as it were, to be told off for this specific purpose, and thus peculiarities in their functioning lead to striking results.

EVIDENCE OF SEGREGATION IN THE HEREDITY OF BUILD.

The best single criterion of Mendelism in any hereditary distribution is segregation. If there is sufficient evidence of segregation in our study of build, then we are justified in concluding that build is inherited in “Mendelian fashion.” There have been several occasions to refer to evidence of segregation in this work; some of this evidence may be brought together here and other added.

1. The difference in variability of the progeny of different matings. The offspring of slender parents are least variable, of fleshy parents most. This is evidence that the fleshy parents carry gametes for slenderness and thus that condition reappears in the offspring; but slender parents rarely carry gametes for fleshiness (p. 56).

2. As a corollary of the above, regression takes place in the progeny of fleshy parents to a markedly greater degree than in the progeny of slender parents (p. 39).

3. The progeny of heterozygous parents are significantly more variable than the progeny of parents belonging respectively to slender and to fleshy stock. This is evidence that the heterozygous parents carry a greater variety of gametes than those of “purer stock.”

4. In different matings of the same type the variability of the progeny differs; apparently, because some parents belong to a special biotype and others simulate the biotype merely through heterozygosity. Thus, if the M × M mating be considered (table X), one finds some families characterized by slight variability of the offspring. The indices of build (English) in such families are given in table 34.

TABLE 34.—_Indices of build of the progeny of M × M matings, in which the parents belong to the M biotype._

+------------+---------------------------------------+------+------+
|Reference. |Indices of build of children. |Mid |Total |
| | |range.|range.|
+------------+---------------------------------------+------+------+
|Bat 2 |32, 33, 33, 35 | 33.5 | 3 |
|Bea 7 |31, 32, 32, 34 | 32.5 | 3 |
|Dic 4 |29, 31, 32, 33 | 32 | 4 |
|Kel 4a |31, 32, M, M, M, 34 | 32.5 | 3 |
|Old 2 |33, 33, 33, 35 | 34 | 2 |
|Rig 1 |30, 32, 32, 34 | 32 | 4 |
|War 3 |30, 30, 30, 31, 32, 32, 32, 33, 33, 36 | 33 | 6 |
+------------+---------------------------------------+------+------+

The slight fluctuation in build of progeny of table 34 is obvious. The modal index of the progeny is close to 33. On the other hand, there are matings of this type which show a much greater variability. These are listed in table 35.

TABLE 35.—_Indices of build of the progeny of M × M matings in which the parents are probably heterozygous._

+------------+------------------------------------------+------+------+
|Reference. |Indices of build of children. |Mid |Total |
| | |range.|range.|
+------------+------------------------------------------+------+------+
|Cas 4 |32, 34, 34, 37, 41, 43 | 37.5 | 11 |
|Coo 5 |31, 33, 35, 36, 36, 36, 37, 37, 38, 38, 41| 36 | 10 |
|Eat A |33, 34, 35, 36, 36, 38, 38, 42 | 37.5 | 9 |
|Fis 2 |33, 35, 37, 37, 37, 39, 40 | 36.5 | 7 |
|Ill 1 |33, 34, 35, 35, 37, 37, 38, 39 | 36 | 6 |
|Mar 22 |31, 32, 33, 36, 36, 40 | 35.5 | 9 |
|Mer 8 |33, 36, 37, 40, 41 | 37 | 8 |
|Nes 1 |32, 33, 33, 37, 41, 51 | 41.5 | 19 |
|Tet A |30, 30, 36, 41, 43, 49, 49 | 39.5 | 1 |
|Rup 1 |30, 30, 31, 31, 33, 36, 38, 40 | 35 | 10 |
|Smi 33 |31, 31, 32, 36, 36, 36, 36, 37 | 34 | 6 |
|Sto 13 |33, 34, 36, 38, 38, 38 | 35.5 | 5 |
|A 01:29 |29, 39, 41 | 35 | 12 |
+------------+------------------------------------------+------+------+

These series of progeny fluctuate in build around 36. It seems probable that the variable progeny are the offspring of heterozygous parents, but this can rarely be proved, since the build of the grandparents is not often available in the present series.

That there are really at least two kinds of matings of the M × M type is shown more conclusively by figure 50, in which two modes, at 33 and 36, respectively, are clearly seen. The conclusion seems to be justified that the mode at 33 is that of progeny derived from the M biotype and the mode at 36 is that of the progeny of heterozygous M parents. The existence of these two modes in figure 50 is thus evidence of segregation.

Additional evidence is found in other matings, likewise. Thus the M × F mating shows two modes, at 33 and at 36 (fig. 51). There is also one other, possibly significant, mode, namely, at 38, in the fleshy group. Since F parents are sometimes homozygous, mating with M would tend to produce medium-fleshy progeny (i. e., 36). The F parents are, however, frequently heterozygous, containing both slender, medium, fleshy, and even very fleshy gametes; and, accordingly, we should expect somatic segregations at about the points 30, 33, 36, and 38 as centers, just where we find them.

Again, the S × M mating yields (fig. 52) modes at 32 and 35 respectively, a shade lower than those of the M × M or M × F matings; but the male offspring (taken alone) gives modes at 33 and 35. Where the mode is at 33 it is probable that an S and an M gamete have met; and this condition is relatively common just because M gametes are common. The mode at 35 may arise from an F gamete carried in a heterozygous M uniting with an S gamete. The S gamete carried in heterozygous M may be responsible for a probable mode which is hidden in the form of a hump at 28 in figure 52.

Attention is called to the probable influence of these modes on the two modes shown in figure 7, at 33 and 35 respectively. These two modes are strong evidence for a segregation in the factors for build. Thus the very irregularity of the polygon of figure 7 is evidence of the presence of segregating factors for build. Moreover, these two modes appear at every age from birth to maturity, as is clearly shown in figure 53. This bimodality is easily accounted for on genetical grounds, but is inexplicable upon the bare nutritional hypothesis of build.

ON THE NUMBER OF FACTORS INVOLVED IN FLESHY BUILD.

It is by no means an easy matter to determine the number of independent factors which are active in the case of a trait that is due to multiple factors. This matter is still less easy if there is only very imperfect dominance approaching intermediacy of the trait in the progeny of the F₁ mating. Yet that is the situation that has to be met in the study of heredity of build.

An attempt has been made in this paper to test the relative probability that there are only two, on the one hand, or three or more independent factors, on the other, involved. We have seen, in each of the various matings, that the result is, on the whole, more closely in accord with the 3-factor hypothesis than that of 2 factors. But this is not to insist that never more than 3 independent factors are involved in build, or never less than 3.

An attempt was made to apply Dr. Sewall Wright’s formula (Castle, 1922, p. 22). Using the data of tables 12, 27, and 31, the value for _n_ (the number of factors) was calculated by the formula

_N_ = _D_² / 8(σ₂²-σ₁²)

where _D_ is the difference between the means of parental pure races, σ₁ is the standard deviation of F₁ and σ₂ is standard deviation of F₂.

Substitution for the letters of their values given _n_ = (39.20-28.50)² ÷8[(6.78)²-(5.02)²] = 0.69. We can only conclude that, owing to small numbers, this formula is not applicable here.

Returning, then, to our conclusion that the assumption of at least 3 gametic factors for some, probably most, cases of build accords best with the results of the various matings, we may inquire what is the nature of these factors. It seems probable that two of them may correspond to dystrophies of the pituitary and thyroid glands respectively and the other or others to other regulatory mechanisms of metabolism in the organism. It may well be that one of our factors is the same as that which causes hypothyroidism; another as that which causes hypopituitarism, and the third that which inhibits the normal development of the genitalia, such as we see functioning in dystrophia adiposo-genitalis. This is, indeed, speculation. However, the speculation is an attempt to give a concrete form to the ideas of multiple factors in build.

If we assume three independent factors for build, it does not follow that these are present in all families. Just as it has been demonstrated that in different strains of mice there are in some cases 3, in others only 2 factors for susceptibility to cancer, so, apparently, it is with build. In certain families the distribution of build in the progeny is best accounted for on the hypothesis that, in such families, there are only two gametic factors for build. In other families there appears to be only one gametic factor for build.

It has been repeatedly pointed out that in many cases where multiple factors are invoked to explain the genetical results, the results can be equally well explained on the theory of blended inheritance. The advantage of the factorial theory is that it brings under one type of heredity even these most difficult cases of apparently blending inheritance. More than that, however, the theory of multiple factors explains a number of phenomena that the theory of blending inheritance does not, such as the greater variability of the F₂ as contrasted with the F₁ generation, and the skewness in distribution of progeny in the back-cross. Thus an F × S mating should produce a “medium” progeny with symmetrical distribution of deviating types on the theory of blending inheritance; actually it produces not only mediums, but an excess of fleshy over slender, which is readily accounted for on the hypothesis that fleshiness is partially dominant over slenderness. In this and other ways, especially in focussing attention on gametes, the doctrine of multiple factors is essential to an adequate interpretation of the results of mating between persons of dissimilar build.

D. SUMMARY OF CONCLUSIONS.

1. Two types of variation in build are to be distinguished: (1) the change in average build that accompanies ontogeny and (2) variations in adult build.

2. The popular idea of build is best expressed as the ratio of transverse chest diameter to stature, or (since the chest diameter is rarely known) chest-girth to stature. When only weight and stature are known, the closest approximation to the chest÷stature ratio is given by the weight÷(stature)² ratio, and this is taken as the standard index of build.

3. The correlation between the standard index of build and the relative chest-girth is, for males, about 0.45.

4. The index of build of adult males is slightly greater than of females, because of the relatively greater chest-girth of males. The average index of build for males is 2.52 (35.8) and for females 2.43 (34.5).

5. There are marked racial differences in build; but they are not so great as the differences in ontogenetic stages.

6. There are geographical differences in build; the heavy build of northern peoples may be due to a physiological reaction or, in part, to a selective survival of the fleshier individuals or strains.

7. The ontogenetic curve of build, expressed by relative chest-girth, is expressed by figure 1 and (for infancy) figure 2. It shows that at birth chest-girth is about two-thirds of stature and diminishes in the male to the age of 12 years; thereafter, on the average, it rises to complete maturity.

8. Build declines temporarily during the first month of life, owing to the physiological difficulties attending adjustment to new conditions. It declines temporarily, again, at about 8 months, probably due to the cutting of the incisor teeth.

9. The heavy build of the infant corresponds to that of the short-legged anthropoids. The long-legged, slender-build stage of the boy of 12 years persists in the Nilotic negroes and many low-grade feeble-minded.

10. In adult life the changes in build vary with families. In those characterized by slender build there is typically little change. In those characterized by fleshy build there is typically progressive increase in weight to 50 years. In some families weight fluctuates greatly at different periods of adult life. In general, though with numerous exceptions, a fleshy adult build is foreshadowed in plump build in childhood.

11. Mass studies on adult build give a polygon of distribution which is skew, the mode being toward the slenderer end of the polygon. There is evidence of more than one mode, and hence that there are two or more types of build. For purposes of description five classes of build are recognized—very slender, slender, medium, fleshy, and very fleshy.

12. The diseases associated with very slender and slender build are: tuberculosis, pneumonia, “nervousness,” melancholia. The diseases associated with very fleshy or fleshy build are: diabetes, nephritis and dropsy, apoplexy, and arterio-sclerosis and paralysis accompanying it; also numerous diseases of the alimentary tract.

13. Fleshy parents have, on the average, in our data, larger families than slender parents.

14. Regression toward mediocrity is less striking in the offspring of slender than of fleshy parents, suggesting that fleshy parents carry not only genes for fleshiness but also for slenderness, while slender parents more rarely carry genes for fleshiness.

15. The offspring of two fleshy parents are twice as variable as those of slender parents.

16. The hypothesis is indicated that genetically build is controlled by multiple factors, with fleshiness tending slightly to dominate over slenderness.

17. There is a marked tendency for persons of similar build (or with potentialities for such) to intermarry. Dissimilar builds are selected against.

18. Two slender parents of slender stock have rarely any progeny whose index of build exceeds 2.2, or just above the upper limit of the slender group. In general, the progeny of slender parents are relatively slightly variable.

19. The slender parents are apparently of two kinds: those carrying only one kind of factor for fleshy build and the other two such gametic factors. The progeny of the former are very slightly variable; those of the latter more variable.

20. By comparing coefficients of variability instead of indices, and by making allowance for the greater range of the fleshy class than the slender class, it still appears that the progeny of fleshy parents are relatively and per unit range more variable than the progeny of slender parents. Absolutely the offspring of the fleshy parents are, as stated, twice as variable as of slender parents.

21. The matings of slender parents and fleshy parents of fleshy stock yield a variable progeny, such as is typically found in the F₁ generation when multiple factors are concerned. The variability is between that of the progeny of slender matings and that of the progeny of fleshy matings.

22. The F₁ generation has not a mode of build that is intermediate between that of the parental stocks, but one that is about the same as of the fleshy matings. This is evidence of partial dominance of fleshiness.

23. The mating of two heterozygous (F₁) parents produces, in general, an exceptionally variable progeny. This is one of the strongest evidences of the presence of genetical factors in build.

24. Many, if not most, parents of medium build belong to biotypes with 1 or 2 independent factors for build. Consequently the progeny of medium × medium matings is often strikingly invariable.

25. The “back-cross” of a heterozygous parent with a slender parent, on the one hand, or with a fleshy parent of fleshy stock, on the other, gives a variable progeny, whose mode is shifted toward fleshiness; but which shows a tendency to fall again into the slender (or fleshy) and medium-fleshy grandparental types.

26. Variations in build are not to be accounted for merely by variations in intake and out-go of calories, but also by the endogenous factors that determine the “economy of nutrition” or the cost in energy of adding an additional kilogram of weight to the body. The factors involved in producing differences in these respects are hereditary factors.

27. The hereditary factors probably work through the intermediacy of special organs that influence metabolism, notably the endocrine glands. The latter thus intermediate between the chromosomal constitution, on the one hand, and control of metabolic processes, on the other.

28. The number of factors involved in very fleshy build is, in some cases at least, three independent ones. But the number is probably not the same in all biotypes. Probably in some there is only one, in others four or more.

29. The factors probably influence the functioning of the thyroid, pituitary, and perhaps other glands.