|
Letter
to Godfrey Barker A rebuttal to Godfrey & Barker, October 1993 |
|
KM Godfrey
and DJP Barker
I was interested to read the article you had published in the BMJ of 14th August on 'Relation of fingerprints and shape of the palm to fetal growth and adult blood pressure' and also to read your reply to my letter published in the BMJ of October 16th. However, I would like to take issue with some of the points you have raised there, particularly since the text of my letter was edited such that the emphasis was changed in the points I have to make. I am glad to see from your reply that the sample of subjects in your study were all Caucasian, though to be fair this was not explicitly stated in your original report. However, it still does not seem that you considered the fingerprint distributions of the men and women separately, a very important point for your thesis given that whorl distributions vary between men and women. Indeed, in none of the tables given in the original paper are the figures separated out for men and women for any of the data. Moreover, I cannot see any figure which states the number of men and the number of women who took part in the study - just the total of 139 (which in itself is hardly a large sample on which to base such far reaching conclusions) which suggests that the sample was not differentiated on the basis of sex for the dermatoglyphic analysis (though it was for the measurement of systolic pressure). Dermatoglyphic protocol, as it is suggested by Cummins, Penrose, Alter et al insists that all dermatoglyphic studies should take the sex of the sample into consideration, hence the study should have looked at correlations for men and women separately. Since whorls are 4-5% more common in men than in women it is obvious if the sample was predominantly male, there would be more whorls in the sample than if the group had been predominantly female. I'm glad also to hear that the fingerprints were classified by an expert from Scotland Yard. However, I am not given any confidence by your assertion that the fingerprints found in the sample were classified into 'the twelve main types of pattern'. As you would surely know if you read FR Cherrill, there are not twelve main types of patterns. It is hard to see how the categories of 'Scarred' and 'Missing' can be seen as fingerprint patterns at all! Moreover, three of the other pattern types which are routinely used by the police, namely 'Lateral Pockets', 'Composites' and 'Accidentals' are so exceedingly rare it is hard to call these a pattern 'type' at all. This leaves us with seven patterns which have a sufficient percentage occurrence and distinctive papilliary formation to be considered as 'types'. However be that as it may, the next stage you present is to 'group' these pattern types together. What you do not seem to be aware of is that these fingerprint patterns are grouped together differently by different people according to the uses to which they wish to put their dermatoglyphic considerations. Certainly it is inexcusable for scientists looking for medical indications within the hand to employ the same 'grouping' classifications as used by forensic scientsts, as their aims are entirely different - they are only trying to recognise different fingerprint patterns for purposes of identification. That this is important is clear when we reflect that all the major fingerprint patterns have been clearly associated with different medical or genetic conditions over the last seventy years. The ususal procedure taken by medical dermatoglyphicists (and I presume this is the procedure you followed as well) is to class both radial and ulna loops together, compounds, twinned loops and whorls together and arches and tented arches together (though it must be said, sometimes tented arches get put together with loops since they do in fact have one triradius). Such a grouping clearly distorts the apparent distributions of the fingerprints - particularly in the case of the whorl. Ordinarily, whorls comprise about 20% of all fingerprints in English males; but when you add in compounds (4%) and twinned loops (4%) you end up with a figure of 28%. I think now you will see my point about how such grouping can 'distort' the results. If you have classified twinned loops and compounds as whorls, then there will appear to have been more whorls in this sample than there actually were and hence the correlations you have made may not be as statistically significant as you think. Moreover, if you have grouped the fingerprint patterns in this way, then you will not have found a relevant medical correlation for whorls alone. In any case, you do not state in your paper what you have done about compounds and twinned loops nor have you investigated what the results would be if you considered the whorl pattern alone. The evidence to show that your figures are somewhat distortive can clearly be seen when we compare the incidence of whorls on the digits in your sample as compared to the incidence of whorls on each of the digits as indicated in Scotland Yard statistics. Your figures are consistently higher - in some cases up to ten or seventeen percent higher - with the exception of the right hand middle finger. Godfrey/Barker Scotland Yard Right hand:
Left Hand: Although any sample of fingerprint patterns are going to exhibit some variations, if you compare the figures, it is obvious that in the majority of these finger distributions, the bulk of the differences between these figures is made up by the addition of compounds and twinned loops. A point worth noting perhaps is that whorls on the middle finger are known to be more common in the criminal population (from whence the Scotland Yard statistics) as compared to the non-criminal population, which would account for the increased incidence in the Scotland Yard sample. Since a whorl is defined as being a different pattern type from the compound and the twinned loop, any conclusions you want to make about the whorl should be about the whorl alone. If you admit that there are 'twelve' main types of fingerprint patterns (ie that the whorl is a distinct pattern from the compound and the twinned loop) then it is illogical assess your results by refering to these three pattern types as if they were one pattern only! Since you have grouped together three different fingerprint pattern types in your study, the most you could conclude is that 'compounds, twinned loops and whorls all together show an association with increased systolic pressure'. Now it may be that the particular group of people you had as your sample had a much larger incidence of whorls on the fingertips than is usual in the English population, though the numbers you have given suggest a figure of about 26.5% which is about right for an English population (if you count twinned loops and compounds as whorls and average male (28%) and female (24%) whorl distributions). But if you really want to test your hypothesis, it would seem advisable for you to conduct further studies in populations where the incidence of whorls is known to be exceptionally high and see if they too have correlative increases in systolic blood pressure. Such studies could include Australian Aboriginals, Tibetans, Dalmatian Coast Islanders, the Eskimos of Greenland, Jewish populations, Chinese and Japanese... all of whom are known to have vastly increased incidences of whorls, often between 40 and 50% - and in the case of Greenland Eskimos up to 70%! But are these populations also known for their incidence of high blood pressure? Since these figures are in some cases more than double the incidence of whorls in the English population and in the figures in your studies, according to your stated hypothesis that systolic pressure increases with the numbers of whorls, these people should consistently have a much higher level of systolic blood pressure overall.
Your view that whorls are 'indelible markers of impaired fetal development' is very largely determined by your insistence that fingerprints are not primarily determined by genetics; it is to this issue we should now turn. Throughout your paper you set your stall out against the genetic hypothesis for the origination of fingerprints, drawing particularly on the evidence from the finger and palmprints of dichorionic monzygotic twins and the researches done on the hands of those suffering congenital rubella in utero. Admittedly, these studies do throw some of the assertions of most dermatoglyphic researchers, since the majority of dermatoglyphicists believe that fingerprints are primarily determined by some mechanism of genetics. However, most also agree that the environment in utero will also have some effect on the developing dermatoglyphic patterns, though few are clear as to how this might happen. But this does not prevent us from viewing fingerprint patterns as having a primarily genetic origin - and indeed, the bulk of the research this last seventy odd years overwhelmingly points in this direction. Some of the studies that have been conducted readily suggest this, particularly of course all the studies into Down's Syndrome, Edwards's Syndrome, Patau's Syndrome, Turner's Syndrome, Kleinefelter's Syndrome and other chromosomal disorders, all of which have found significant dermatoglyphic deviations which can be clearly associated with chromosomal and hence genetic disorders. Since most of the major dermatoglyphic researchers of this century (to name LS Penrose, H Cummins and S Holt as but a few) have conducted and verified such findings, it seems you are either unaware of the sum total of their work or that you simply wish to ignore the substantial amount of evidence in favour of a genetic origination of fingerprints. Moreover, specific studies have been conducted which have specifically reveal a hereditary component in dermatoglyphic distributions - not least LS Penrose in his studies of the hereditability of the location of the axial triradius in Down's Sydrome and Sarah Holt in her pioneering studies which resulted in a quantifiable means of determining twin zygosity on the basis of fingerprint patterns alone (identical twins having a .95 correlation in Total Finger Ridge Counts). This is not to mention the hundreds of studies done by researches of all medical persuasions who have found the different fingerprints to be associated with a genetic disposition towards different conditions - eg Sank et al: tented arches and schizophrenia, Gottlieb and Schuster: simple arches and constipation, Weinreb: ulna loops and Alzheimer's disease - and the studies done by researchers like TJ David into families with rare and unusual fingerprint patterns (Ridges off the End Syndrome) which have shown a genetic basis to dermatoglyphics. All in all, you seem either to have overlooked all this research or else we must conclude that you are entirely ignorant of it all. This is all before we even begin to consider all the anthropological research that has been done! All this research has shown that fingerprint patterns vary enormously from race to race and from people to people, as we have already seen. Since fingerprint pattern distributions are clearly racially based, this is overwhelming evidence that fingerprints have at their root a primarily genetic basis. (Or will you now also propose an argument to the effect that racial characteristics have no basis in genetics either?) As we have seen, one reason you seem to see whorls as 'abnormal' is because of your adherance to an adverse in utero environmental causation for the formation of fingerprints, which you seek to present in favour of the genetic argument for the formation of fingerprints. However, you have only cited a few small studies that in no way counter the balance of the evidence which is in favour of a genetic account for the origin of fingerprints. To make a short philosophical aside, we can concur with the empiricist philosopher David Hume when he observes that knowledge can never amount to a certainty and that at most, we only ever have the balance of probabilities. The reasonable man accepts the hypothesis that has the balance of evidence in favour of it. In preference to the genetic account for the formation of fingerprints, you turn instead to the highly implausible theory of Mulvihill and Smith that whorl fingerprint patterns are formed by oedemic swelling on the tips of the fingers in utero. According to their theory, large swellings of the fingertips will produce whorls whereas little or no swelling will produce simple arches. Swellings which are laterally displaced will produce ulnar loops or radial loops, depending on which side of the finger the swelling occurs. You state that: "Mathematical considerations dictate that parallel ridges stretching over a more bulbous finger pad will form whorls rather than simpler arch or loop patterns" and that: "Patterns are formed like contour lines around a hill with the centre of the pattern corresponding to the peak of the hill". But if you actually look at the shapes of fingerprint patterns closely, it is hard to see how this could be true for all fingerprint pattern types. Note that Mulvihill and Smith only utilise their hypothesis for four of the seven/ten/'twelve' types of fingerprint patterns. What about the other patterns, the more complex patterns such as nutant loops, twinned loops, composites, accidentals, lateral pockets or even tented arches for that matter? Could these patterns really be formed in this way? It is easy to use the analogy of hill contours as far as the whorl is concerned.... but lets think about this for a moment. Surely if swelling on the fingertips causes the formation of whorl patterns, then 'slight swelling' of the fingertips will cause little whorls, rather than simple arches, and laterally displaced 'swelling' will cause laterally displaced whorls, rather than radial or ulnar loops. In support of your hypothesis, you cite research conducted on the dermatoglyphics of those suffering from Turner's syndrome, where oedemic conditions are often noted and are thought to affect the fingerprints (though this does not seem to be certain if we read the Penrose article cited in your original article). However, the biggest problem with this is the fact that various studies have shown that the fingerprint pattern distributions for those suffering from Turner's syndrome is not significantly different from the fingerprint patterns of normals! And yet, if the theory that 'abnormal' fingerprint patterns were indeed caused by abnormal swelling on the fingertips were true, we should find reports of vast numbers of whorl patterns in the hands of those suffering Turner's syndrome... It is interesting to observe also that although those suffering rubella embryopathy in utero are known to have large incidences of whorls (incidences of up to 50% have been reported), oedemic swelling of the fingertips is not especially noted in this condition and you fail to mention if this is the case (or not) in your paper. Surely if oedemic swelling occurs in rubella embryopathy it would serve your case much better than Turner's syndrome? From this basis, you then go on to assert that the higher prevalence of whorls on the right hand over the left can be accounted for by some anatomical differences in the origins of the arteries that supply blood to the head and arms of the developing foetus. However, your description rather rests on the presence of adverse conditions in utero which means more blood flows down the right subclavian artery rather than the left, hence: "whorls on the right hand may result from selective preservation of brain growth". In any case, your theory is that abnormal conditions in utero produce a different distribution of fingerprints on the two hands. The palpable untruth of this theory, that whorl fingerprints are formed by some accident of blood supply to the right side of the body, is readily revealed when we remember that whorls are more common in men than in women and are more common in Chinese and Japanese peoples than they are in Caucasians. Presumably then, women, Chinese and Japanese people have completely different blood supply arrangements than English men! If you are saying this distribution of whorls occurs because of some 'abnormal' conditions of fetal growth, then this is tantamount to saying that all those races of peoples with large numbers of whorls experienced abnormal intrauterine conditions. Whilst nobody is denying that intrauterine conditions can have an effect on the development of dermatoglyphics, it seems clear that you have not really offered a coherent and viable alternative to the genetic hypothesis for the origination of fingerprint patterns. Moreover, you are quite wrong to assert: "The statement that fingerprints are genetically determined reflects the Cheirological Society's unproved beliefs". This statement quite clearly shows a paucity of research as to what the Cheirological Society does or does not believe about the origin of fingerprint patterns - as well you would know if you had read any of our literature or knew anything about the nature of our work. My arguments against your hypothesis have been from entirely within the corpus of existing and published scientific researches into the origin and significance of dermatoglyphic patterns. Far from being an 'unproven belief', the genetic hypothesis for the origin of fingerprint patterns has been, and still is, widely held in the scientific community that cares to consider these things. Even a cursory glance at the extant literature on the subject will tell you that the majority of researchers in the field consider the fingerprint patterns to be primarily genetic in their origin. Why else would there have been such extensive research into chromosomal anomalies, congenital heart defects and other genetically based disorders? Why else would papers appear with titles such as 'Dermatoglyphics in Medical Genetics' or 'Dermatogyphics and Human Genetics'? Why else would Sarah Holt entitle her book 'The Genetics of Dermal Ridges'? As I have tried to show throughout, the weight of evidence for the genetic hypothesis for the origination of fingerprint patterns is considerable and is not a randomly held 'unproven belief'. Finally, to go back to methodological criticism for a moment, once again we find evidence of your lack of familiarity with dermatoglyphic precedent and procedure. The fact that you have misused the 'ATD' angle measurement only serves to suggest, once again, that you are not really very familiar with the wok that has been done in this field. Unless, once again, you have chosen to ignore the work that has gone before you. Either way, that does not inspire confidence in your research. As I said in my letter, your abuse of this measurement comes from the fact that the ATD angle was originally established by LS Penrose as a means of quantifying the location of the triradius at the base of the palm and was never intended as a measure of palm size, either by length or breadth. Penrose created this formula precisely because the location of the axial triradius can vary enormously from one hand to the next. However, I note your concession that you did not take displacement of the axial triradius into consideration in your study and that this might have produced some error in the estimation of hand shape. Nevertheless, this is still an admission that you didn't know what the ATD angle measurement was for - since if you had known this, you would have taken it into account in your study or wopuld not have used the ATD angle to measure what you did! Of course, this does not mean that there is no correlation between hand shape and blood pressure. But this would have perhaps been more acurately ascertained if you had used a different method of measuring the shape of the palm - for example, just using a ruler to measure the length and width of the palm. But since you have used an inaccurate measuring tool, in all fairness, you canot acknowledge this and then say 'oh yes but it doesn't really matter and won't make any difference'. How do you know? A more scientific response would be to say: 'Oh well, we can't be sure of that correlation until we have measured it all again using a different technique'. If you admit that your measuring device was faulty or possibly inaccurate, then you are admitting imprecision and you can't rest content with the measurements you have made. The biggest irony of this though is that it is the location of the axial triradius and the measurement of the ATD angle which is perhaps the most certain of all the dermatoglyphic indicators that skin ridge patterns have as their origin a primarily genetic cause. The displaced axial triradius is extremely common in all the major chromosomal disorders, including Downs Edwards and Pataus syndromes as well as the less common conditions such as Trisomy 8, Cri du chat, Wolf-Hirschhorn syndrome and others. In addition to its significant association with genetic/chromosomal disorders, in fact, Penrose found its displaced location to be the most important indicator of a genetic predisposition to producing a Down's Syndrome child. His researches showed that the location of this pattern was largely determined by genetic factors and that from its presence in the hand of the mother, he could ascertain whether she was likely to produce a Down' Syndrome child. Since one of the conclusions of this study is simply that the presence of at least one whorl is an indicator of high blood pressure, this is a hypothesis that can be readily tested: the more whorls you have the higher the blood pressure. But since it is quite normal to have one whorl - even in the sample presented here 93/139 (67%) had a least one whorl - should we reasonably conclude from this that the majority of the English population has high blood pressure? It is for all these reasons that I am rather sceptical about the validity of this research, both in terms of methodology and in terms of the results obtained. Since much of the earlier dermatoglyphic literature has been completely ignored, I can only conclude that the results of this research are at the very least questionable, if not completely doubtful. Although I cannot comment on other aspects of the study, such as head size and placental weight at birth, I note the statistical criticisms offered from the University College of Wales and other obvious facts like the small size of the sample; but, moreover, the rather curious findings, which seem totally at odds with perceived wisdom, that smoking and body mass showed no correlation to blood pressure levels at all! yours sincerely,
|
| © Christopher Jones 1989-2012 |