Wikipedia:Reference desk/Archives/Science/2013 December 3

Welcome to the Wikipedia Science Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

December 3

Skin color and auto accidents

According to a report by the Centers for Disease Control and Prevention, the death rate for American pedestrians in auto accidents, 2001-2010, was much higher for blacks than for whites. This report was the only solid source I could find that's at all relevant to skin color, so I'll have to ask for help finding something that interprets the data. Is anyone aware of studies that directly address the effect of skin color on rates of auto-pedestrian collisions? I was left wondering about skin color itself while walking along a street after nightfall this evening: I'm left wondering if my light-skinned face and hands are more visible to a driver than they would be if my skin were dark. Since statistically average white and black Americans differ in much more than just skin color, we obviously can't say that the different death rates are purely the result of drivers seeing white pedestrians more easily than black pedestrians. Has anyone done a study that seeks to control for all other factors (e.g. controlling for percentages of people of each race walking in places without sidewalks, walking at night, walking while drunk, etc.), leaving only the visibility of the skin as a variable? Lest this be misinterpreted: I'm not fishing for anything, nor suggesting anything about common black culture versus common white culture; I would ask the same question if common concepts of race relied primarily on eye color and ignored skin color entirely. It's purely a question of the relationship between melanin and optics. Nyttend (talk) 02:14, 3 December 2013 (UTC)

• Seriously? (Are you from the Alps? Methinks thou dost protest too much.) Blacks in the US are overrepresented as part of the urban population, where car-pedestrian impacts are most common. μηδείς (talk) 03:43, 3 December 2013 (UTC)

I think you'd definitely need a controlled study under strictly defined conditions to rule out all of the confounding variables such as differences in fashion. I don't think you can take actual incidents and control every factor like that, because for example how would you categorize every garment every person was wearing? (I know there are definitely some colors that are more popular with people of a particular race in a given area) Wnt (talk) 03:39, 3 December 2013 (UTC)

<sigh> Recite ten times: "Correlation does not imply causation" (then read our fine article: Correlation does not imply causation).

It is true in the US that (for example) on the average, (per Racial wage gap in the United States) black people only have 65% of the income of an average American. So perhaps the truth is really: "Poor people have a higher death rate than the middle classes in pedestrian auto accidents". Perhaps they live in neighborhoods with worse street lighting and the real news is "People who live in well-lit neighborhoods have a lower death rate in pedestrian auto accidents"...which takes us from a socially divisive message to a socially useful one that says "We should improve street lighting in the poorer parts of America"....or not. We don't know because the science wasn't done right.

For all scientific studies, you absolutely must employ at least one control group - and for this study, you'd need MANY control groups before jumping to any kind of conclusion whatever. So in this case, one would want to control for income by (say) comparing the death rates for people of different skin colors in identical income groups - and who live in identical parts of the city, etc, etc.

Jumping straight to the conclusion that skin color is the "cause" is a very, very bad idea when all you have is a single "correlation".

So this study (as it stands) is entirely useless and may safely be filed away under the heading of "unqualified data" until we see some halfway reasonable study about why pedestrian accidents happen.

SteveBaker (talk) 03:40, 3 December 2013 (UTC)

Have you read any of the study, or even the question, before denouncing the study as "entirely useless"? The study makes absolutely no claim about why black people have more accidents. It simply lists accident rates by race, gender, urbanization level, age, and other factors, and doesn't even put any emphasis on race.

If you read the question, you'd know the OP asked "Since statistically average white and black Americans differ in much more than just skin color, we obviously can't say that the different death rates are purely the result of drivers seeing white pedestrians more easily than black pedestrians. Has anyone done a study that seeks to control for all other factors..." He's asking precisely for a controlled study. --Bowlhover (talk) 04:36, 3 December 2013 (UTC)

The first thing missing that I see is a comparison to other accidental death rates and non-accidental death rates. Considering that it's only 13% of vehicle related deaths, the next study I would look at is wheher vehicle/vehicle accidental deaths are correlated in the same way as pedestrian/vehicle deaths (that seems surprisingly missing). Those studies should be available broken down by the same categories. The next question is whether pedestrian/auto deaths are a significant cause of accidental death in each group or is it in the noise (this is particularly egregious omission as it appears to be raising alarm bells for an aging population - death rates increase by age but do other accidental death rates, such as falling, increase faster? We don't know.) . Accidental death rates by cause are available as well. As a whole, the study starts with auto accidents being overall #1 cause (but not pedestrian/auto) of accidental deaths. Then makes a scary number for elderly pedestrian death rates but fails to rank it for the elderly group. That's pretty poor. Also, as it states, the study doesn't control for alcohol but uses cause of death on death certificates. "Vehicular assault" is not accidental and may vary by geography as to whether it's accidental or not. The study clearly indicates that there is a correlation to urbanization and that within those areas, the other factors correlate to each other. "Approximately three fourths of all pedestrian deaths in 2010 occurred in urban areas (3). Higher pedestrian death rates in urban areas are, at least in part, a result of more concentrated vehicle and pedestrian activity in these areas. The current study found that for many age groups and racial/ethnic populations, patterns in pedestrian death rates by level of urbanization were similar to those for overall pedestrian death rates" which seems to mean that for 3/4ths of all pedestrian deaths, race/ethnicity was not a factor and the differences appear when non-metro areas are factored in. --DHeyward (talk) 14:06, 3 December 2013 (UTC)

My first thought was the same as Medeis, i.e. it could have to do with being relatively more clustered into urban as opposed to rural populations. In the latter, you almost have to try to get hit. ←Baseball Bugs ^{What's up, Doc?} carrots→ 04:51, 3 December 2013 (UTC)

One might also check whether the death rate mentioned covers deaths in pedestrian–car collisions only or also caused by such collisions, and then consider the time and quality of medical service available to those injured. --CiaPan (talk) 06:36, 3 December 2013 (UTC)

Some thoughts:

1) As stated before, blacks tends to live in high density communities where car-pedestrian accidents are more common.

2) Also, blacks who live in such places are less likely to own a car, so more likely to be a pedestrian.

3) I have noticed, however, that there seems to be a black cultural habit of walking down the middle of the street, versus on the sidewalks. This seems to be particularly true of male teens. I even posted a question about that once on the Ref Desk.

4) I've also noticed a total lack of reflective clothing, although I suppose being able to hide from those who wish to do you harm, like the police, in some cases, might explain this. StuRat (talk) 07:53, 3 December 2013 (UTC)

The intro states that the highest death rates are native americans and those > 75 y/o. There needs to be lots of controls before assigning causative vs. correlative. In addition, there is no analysis of whether the differences are statistically significant. The study starts off with about 33,000 vehicle deaths and then states about 13% are pedestrian related. The biggest glaring hole that I see is that it fails to compare accidental deaths of pedestrians against other accidental deaths (i.e. trips/falls, drowning, etc). It's pretty bold to say that the aging population may suffer more pedestrian/auto deaths if it's dwarfed by trip/fall deaths or auto/auto deaths. Even though as a whole, auto/auto deaths are the leading cause of accidental death (and it doesn't even state whether they consider alcohol related deaths to be accidental or vehicular assault), the study doesn't seem to break down whether pedestrian/auto deaths are, as a whole, significant. Nor does it break down each group and identify whether pedestrian/auto deaths as a cause of accidental death is higher or lower than other causes. It's number 1 overall but even though it's a higher death rate for > 75 y/o's, it could be the 5th highest cause of accidental death for > 75 and trending lower as the cause of accidental death even while the rate goes up. There's a lot of inferences made that don't seem to have a rigorous mathematical underpinning of sampling statistics and correlations. --DHeyward (talk) 13:29, 3 December 2013 (UTC)

For people wearing dark clothing at night, against a dark background, it would make a difference for the distance at which one is first noticed. Weber's law and Ricco's law would be relevant. Weber's ratio seems to have the best correlation with detection distance according to this. Skin reflectance for different populations: African (Black) 5 - 10%, African (Pigmy) 10 - 15%, Indian (India) 15 - 30%, Iranian 20 - 40%, European 35 - 60%, source: Jablonski, N.G. and Chaplin, G., The evolution of human skin coloration. J Hum Evol 2000; 39: 57-106, quoted on a photography forum. Pictures like this or this suggest that when dressed in dark non-reflective clothing at night, the skin would be the most (or only) visible feature. In those circumstances white people would be detected at a greater distance than black people. And (available) time to react depends on detection distance. Same would apply to blond versus black hair when seen from the back. That doesn't mean of course that the difference in death rates are caused by skin color, we don't even know when these accidents happen, but the idea that skin color can affect visibility and therefore the risk at night seems quite reasonable. I doubt however that anyone has studied this, might be controversial... Ssscienccce (talk) 14:08, 3 December 2013 (UTC)

Just because something is controversial doesn't mean it hasn't been statistically studied. For example, the U.S. Army and the Marine Corps have determined beyond a reasonable doubt that these skin colors are the most difficult to see. Sort of a pixelated grayish green with brown and olive patches to break up the outlines... these patterns are not arbitrary; they are the result of a century of statistical testing on human perception. I'm sure we can find hundreds of papers listing variations on those experiments. In some cases, bright high-contrast colors and geometric shapes are more difficult to see. In any case, I think we can thoroughly discredit the idea that visibility due to skin coloration is the determining factor in automobile-pedestrian collisions. The studies that Nyttend actually linked to implore further research into other sociological factors to provide new explanatory variables. Nimur (talk) 14:21, 3 December 2013 (UTC)

More than once while driving at night, I've nearly hit white guys wearing dark clothes and not paying sufficient attention to traffic. Skin color isn't much of an issue, as you only see a small portion of their bodies. It's the clothing that's the problem. That's what a study should look at. ←Baseball Bugs ^{What's up, Doc?} carrots→ 15:13, 3 December 2013 (UTC)

I meant studies specifically looking at pedestrians being killed due to low visibility of their skin color: inadequate street lighting will be more prevalent in poor neighborhoods, so a study telling them their skin color is "to blame" might seem a bit cynical... Ssscienccce (talk) 16:58, 3 December 2013 (UTC)

That's another "obvious" claim with no basis in reality. The most well lit neighbourhoods I have ever lived in were black and Latino neighbourhoods in NYC and urban areas of New Jersey with low income residents. The most poorly lit were upper middle-class ones in rural-suburban South Jersey and at the Jersey shore. μηδείς (talk) 03:16, 4 December 2013 (UTC)

And I'm saying that for a given person, there's a lot more clothing visible than skin. So clothing is more likely to be a determining factor. Obviously, if a black person is wearing dark clothing, their odds of being seen are liable to decrease. But I'm also telling you, from experience, that a white person wearing dark clothes is also difficult to see. ←Baseball Bugs ^{What's up, Doc?} carrots→ 18:49, 3 December 2013 (UTC)

I was responding to Nimur, not you. And of course clothing is much more important, that's why every autumn we see safety campaigns with people dressed like glowstick adverts on tv. I didn't say white persons in black clothes are easy to see at night, I'm saying that skin color will make a difference in visibility level in some situations. I didn't say anything about it's relative importance. Ssscienccce (talk) 22:12, 3 December 2013 (UTC)

And that's why I always get off the street when a car comes along, even though my skin is light, since most of my cold-weather clothes are dark. Since rural and suburban residents are more white than urban residents, I figured that the sidewalk factor might be a big issue: most streets in urban settings are accompanied by sidewalks, unlike many streets/roads in suburban and rural settings, and many of those suburban or rural thoroughfares are lit only by the moon and the stars, so place of walking and quality/quantity of street lighting aren't purely a problem for the urban resident. For example, I know a woman whose father was killed while running along a Kansas farm road in the 1970s, because of factors other than visibility; of course that incident might only be anecdotal, but non-visibility factors such as simple inattentiveness would definitely need to be controlled for the kind of study I was envisioning. Nyttend (talk) 01:15, 4 December 2013 (UTC)

How can we characterize the quality of OECD data on working hours vs productivity?

Recently on the Humanities Reference Desk, this relation between hours worked per year and productivity from an article in The Economist was called "low quality." While there do appear to be bifurcated strands along the same trend, I am not sure I would call such a correlogram particularly low quality. What do Science reference deskers say? EllenCT (talk) 08:28, 3 December 2013 (UTC)

The Economist article linked to this paper that found the same result from a totally different data set. The paper formally reviews their data collection and processing, including the statistical significance of their conclusion. It seems that we can't realistically attack this correlation by attributing it to poor data, nor to poor methodology, nor to statistical insignificance. A complete run-down of the statistics in that paper rule out several confounding factors, and make a strong case for the significance of the correlation (which they term the "elasticity" of hours worked versus compensation rate). From a purely statistical point of view we can confidently say that this correlation is strong, it is significant, and it emerges from multiple different data sets. So, if we want to maintain a scientifically valid standpoint, we should seek a causal explanation for the data.

The authors provide four plausible causal hypotheses, and because they are behavioral economists, they have a preference for a behavioral explanation: the workers are choosing to end their days early when they earn well; they have a fixed income target. I am more inclined to believe their liquidity-constraint hypothesis, based on my personal experiences; but the authors provide compelling counter-evidence by showing that high-capital laborers exhibit the same behavior as low-capital laborers.

I think the most critical point might be that productivity, as defined by the OECD and most other economists strictly refers to economic productivity: activity that has monetary value. That may differ from your intuitive notions about productivity. The next time you pull long hours at school or at your job, you might be working very hard, and you might even be making progress toward some end goal. But consider, for a while, whether your activity is productivity as an economist sees it: are you measurably increasing the GDP ?

Nimur (talk) 13:44, 3 December 2013 (UTC)

In that topic someone said "the data quality is likely to be poor", claiming that in his industry the nominal number of hours was 1620 while the actual number was about 3000. According to OECD Measures of Total Hours Worked, "in many countries actual hours are derived from establishment surveys for production and non supervisory workers in employee jobs and from labour force surveys (LFS) for self-employed, managers and supervisory workers, farm workers and public sector employment. Hours lost due to sickness are estimated from the number of days not worked from social security registers and/or health surveys..."

Germany recently revised its system to better account for workers with few hours, France is now excluding short rests and work breaks from the total (which is contrary to the guidelines)...

Labour force surveys seem to overestimate the number of hours worked and underestimate the hours not worked due to holiday, vacation, sickness, maternity leave etc.. the difference with establishment-based surveys was 1 to 3% (in France, Germany, Netherlands and Switzerland). 3% won't make much of a difference to the graph.

The data may not be totally accurate, but it's likely the best data available. Ssscienccce (talk) 20:23, 3 December 2013 (UTC)

2011 Tōhoku earthquake and tsunami.

Was it the largest in Japan? The 869 Sanriku earthquake was believed to could have been around 9.0. Applies to Japan. --78.156.109.166 (talk) 10:20, 3 December 2013 (UTC)

The earthquake was the largest in the vicinity of Japan since recording began. It's impossible to say whether there were larger quakes earlier. Note though that the Tohoku quake had its epicenter well offshore. The shaking on the mainland lasted for a long time but was not all that intense. There have been many quakes that caused more intense shaking at points on the mainland, and more severe shaking-related damage. The damage from the Tohoku quake was mainly due to the tsunami rather than the shaking. Looie496 (talk) 17:07, 3 December 2013 (UTC)

How much damage/casualties from the quake? --78.156.109.166 (talk) 19:52, 4 December 2013 (UTC)

Studies of the 869 Jogan Sanriku earthquake suggest that it was similar in scale to the 2011 event[1][2]. The amount of inundation was found to be similar, the earlier tsunami's extent was underestimated as it became clear that the sandy tsunami deposit from the 2004 tsunami did not represent the full amount of inundation - mud and silt went significantly further. Taking that in to account it looks like the 869 event may have reached 9.0 magnitude, or even larger. Mikenorton (talk) 23:36, 4 December 2013 (UTC)

Sex determination by the father?

Do "normal healthy" human males produce 2 kinds of sperm, one kind that results in male children and another kind that results in female children? (option A)

Or do "normal healthy" human males produce 1 kind of sperm that can result in either male or female children, depending on something else? (option B)

If the answer is A, are there conditions in which some men only produce one kind or the other? In other words, are there some men who can only have sons or only have daughters?

If the answer is B, what are the main things that determine whether child is male or female?

Thank you, CBHA (talk) 18:09, 3 December 2013 (UTC)\

The father is the sex-determinant of the child. It has to do with whether a given sperm cell has the X-chromosome or the Y-chromosome. And of course there are gazillions of sperm cells in a given "batch". I'm sure there is a great deal of info here. Try gamete for a start, and see where it leads. ←Baseball Bugs ^{What's up, Doc?} carrots→ 18:47, 3 December 2013 (UTC)

For the most part, "the sex-determinant of the child" is not so much the father, as it is random chance. Yes, the winning sperm is the one that determines the child's sex, but the father as such has very little influence over which sperm "wins". --Trovatore (talk) 21:16, 3 December 2013 (UTC)

Sure it's random, for the most part. I didn't say the father had a choice. But he's still the biological determinant of the child, as opposed to the mother, who isn't. ←Baseball Bugs ^{What's up, Doc?} carrots→ 00:19, 4 December 2013 (UTC)

Well, no, sorry, that's wrong. The father is not the biological determiner of the child's sex. Which kind of sperm happens to win, that's the determiner, but not the father. --Trovatore (talk) 07:49, 4 December 2013 (UTC)

The problem is a linguistic one. That is, 'determines' means different things to different people. To many it sounds like a synonym for "chooses", as in "I will determine which movie we see". In that interpretation of the word, the father does not "determine" the sex of the offspring. StuRat (talk) 09:47, 4 December 2013 (UTC)

In biology class, they always told us, "the father determines the sex of the offspring". I'll rely on what my biology teacher said, and what internet sources say,[3] as opposed to nitpicking over the meaning of the word "determine". Maybe Il Trovatore would prefer to say that the source of the sex of the offspring is the father's sperm. But it's ultimately the same thing. ←Baseball Bugs ^{What's up, Doc?} carrots→ 23:25, 4 December 2013 (UTC)

Well, science is about being precise, and a scientific statement that can be taken two ways doesn't seem at all precise to me. StuRat (talk) 08:04, 5 December 2013 (UTC)

The sex of the child was once thought to somehow be determined by the mother. Once they discovered chromosomes, they realized it was determined by the father. Maybe they didn't feel like talking about gametes in public, so saying the sex of the child was determined by the father was sufficient. Or maybe they just didn't want to present the whole story every time. How precise does it need to be? It's determined by whichever sperm cell happens to get to the nucleus of the ovum first. And then you get into more and more details. Do you want a lecture? Or do you want to know which parent is the source of the sex of the child? It reminds me of the story about the kid who comes to the parent and asks, "What is sex?" So the parent goes into a long, graphic description of the reproduction process. The young'un, looking like a deer in headlights, finally says, "I have this paper to fill out, and it says, 'Sex __' What should I put there?" ←Baseball Bugs ^{What's up, Doc?} carrots→ 14:37, 5 December 2013 (UTC)

The fact that the child's sex is determined by which sperm wins does not imply that it's determined by the father. In fact, it doesn't even imply that it's not determined by the mother (or more precisely by the environment of the mother's reproductive tract). If Shettles was right, then the pH of the mother's reproductive tract is actually a major contribution to determining the sex of the child.

Now, I gather that Shettles' theories are no longer in favor. Maybe they're even considered to have been refuted; I'm not an expert and I wouldn't know. But it is certainly not as simple as saying the sex is "determined by the father". --Trovatore (talk) 01:54, 7 December 2013 (UTC)

(EC) As described in the Spermatozoon article, human males produce two kinds of sperm cells: some contain a Y chromosome, which will produce a boy, and some contain an X chromosome, which will produce a girl. I can't find any condition that results in a man only producing one of the two types of sperm cells. But see Human sex ratio#Factors affecting sex ratio in humans. Red Act (talk) 18:52, 3 December 2013 (UTC)

Males have two sex chromosomes, an X and a Y. Sperm cells are the result of meiosis, a process involving cell division, where one cell will get the Y chromosome and the other the X chromosome. These cells split again to create four sperm cells, two with X and two with Y. So a man should produce the same number of X and Y sperms. Ssscienccce (talk) 19:28, 3 December 2013 (UTC)

One thing I looked into is whether an aneuploidy involving a lack of an X chromosome could produce a man who could only father boys. However, according to Aneuploidy#Types, having no X chromosome is lethal. Red Act (talk) 20:47, 3 December 2013 (UTC)

Thank you all. Followup question - are there "factors" in the female (chemistry for example) which differentially affect sperm? Ie, that make it more likely that a particular woman would have male children or female children even though the sperm she received had a 1-1 sex ratio? CBHA (talk) 22:08, 3 December 2013 (UTC)

There was a brief splash made back in the Seventies or so (update: actually Sixties) when some researcher claimed he had a method of making a boy or girl more likely, depending on parent preference, based on the pH of the woman's system. He claimed that an acidic environment was hostile to all sperm, but that the X sperm were tougher and more likely to survive it, so an acidic environment favored girls and an alkaline environment boys (because the ratio is not in fact 1–1; it's actually about 2–1 in favor of Y sperm, if I remember correctly). Here it is: Shettles method. It appears that it is not now considered to work. --Trovatore (talk) 22:14, 3 December 2013 (UTC)

And see also Sex selection. The only preimplantation methods which are known to work I'm aware of, and mentioned in our article, are those which involve artificial insemination or in vitro fertilisation. in other words, any method which is supposed to use sexual intercourse for conception has no evidence of working. Note that while controversial, there is a fair amount of interest in sex selection particularly male selection in certain cultures, so there is some degree of research. So for many popular methods, there's a fair chance the is something which has shown the method may not work (or alternatively it may be there's good reason to think it won't work even without research). (You can of course use post conception selection with sex but that's even more controversial and doesn't seem to be what the OP is getting at.)

And there any plenty of supposed natural methods [4] [5], many even more junky than Wheelan or Shettles [6].

Incidentally the sperm sorting methods could provide a clue as to what could potentially work. However the methods known to work and even some that don't appear to [7] are not the sort of thing you can expect to emulate. Of course as with most things, this sort of research gets confusing. As per the prior link, there's some evidence sorting via the Ericsson method doesn't actual achieve any sort of specific sex sperm enrichment, but there's also some evidence (as per our article) it produces a non expected sex ratio at birth. (Sperm sorting is also of strong generally uncontroversial interest for use in non human mammals. While there's no guarantee these methods will cross over and methods which may work in humans may not work for other mammals particularly somewhat more distantly related ones like cows but it's another area of research or relevance.)

P.S. For completeness I should mention I believe there is some evidence the sex ratio varies depending on the health of the mother [8]. This could be pre implantation or post implantation or both. I think most think it's more likely to be predominantly post implantation (with at least some degree perhaps even most being non concious factors). But even if there is some degree of pre implantation causation, I don't think many would seriously suggest starving the mother for a minor change in the sex ratio.

Nil Einne (talk) 23:02, 3 December 2013 (UTC)

Maybe. See Maternal influence on sex determination. Red Act (talk) 23:20, 3 December 2013 (UTC)

There is related information at http://www.boygirldiet.com/index.htm.

—Wavelength (talk) 23:26, 3 December 2013 (UTC)

Very interesting, though I'm not sure that's a WP:RS, let alone authoritative. I make no judgment, just flagging for propriety :) SemanticMantis (talk) 00:41, 4 December 2013 (UTC)

To answer the original question, there are chromosomal karyotypes that result varying chromosomal counts. See Klinefelter syndrome for description of the most common anomaly. I don't know how often that karyotype produces viable sperm or what the M/F ratio might be. --DHeyward (talk) 09:36, 4 December 2013 (UTC)

Best breeding age for human females

After what age does the offspring become likely to be genetically bad? At what age should the femaile be impregnated and by what age male to insure the child is genetically fit? — Preceding unsigned comment added by 74.14.28.186 (talk) 21:15, 3 December 2013 (UTC)

It's a sliding scale for the probability of Down's syndrome, for example. But even young, healthy mothers frequently have miscarriages. There may be an optimal age range for carrying children, but there are no guarantees. ←Baseball Bugs ^{What's up, Doc?} carrots→ 00:22, 4 December 2013 (UTC)

Your heading only says "Best breeding age" while your question only refers to genetics. As Teenage pregnancy says, there are other problems with young mothers. I suspect (don't know) that if we only look at genetics then younger is generally better. Other health factors are more important when you for example go from 20 to 13, so I don't recommend the latter. PrimeHunter (talk) 01:08, 4 December 2013 (UTC)

• Biologically, 20 is best, assuming the female is fully mature. The range of 20-34 is considered best because of fertility, and the older the mother the more financially stable and mentally mature. This is all over the place if one simply googles the question. Of course the answer now also works depending on one's state of mind. μηδείς (talk) 05:14, 5 December 2013 (UTC)

In a sense this isn't a scientific question. Evolution clearly has an impregnate early, impregnate often agenda; 'barefoot and pregnant' was, after all, the human condition for a very long time. A risk of death, even for the mother, was balanced against the gain of offspring. In evolutionary terms the "what if the child is genetically unfit" worry just doesn't exist - either it survives (victory!) or it doesn't, which in cold arithmetic is no loss. Now a person with modern sensibilities can say that the chance of having a healthy child is not worth a certain chance of one with Down Syndrome, but if so... when is it ever worth it? If you truly take a 'do no harm' approach then pregnancy is right out, Shaker style. So the balance you seek is a compromise between philosophies, a value judgment you make at an arbitrary rate of exchange. We might show you incidence graphs of various conditions (I'm sure there must be one in Down syndrome) but you have to make your own decision. Wnt (talk) 14:51, 5 December 2013 (UTC)

Constrictor snakes' skin

We've all seen the discussions of the ways by which constrictor snakes are sometimes able to swallow large prey, e.g. the goat or deer pictured here. But what about their skin? How does the skin stretch to facilitate a temporarily far larger snake? How doesn't it just tear in pieces and kill the snake by blood loss? Force an adult human into children's clothing and you'll up with little bits of torn cloth; since the snakes eat much too fast for the skin simply to grow big enough to accommodate the large prey, I don't see the difference. After all, a hungry constrictor snake doesn't have massive amounts of extra skin, which one might initially imagine as the expansion mechanism. Nyttend (talk) 23:17, 3 December 2013 (UTC)

Here's a start, it's all about the collagen structures. Here's a paper "Mechanical behaviour of snake skin", that says

“ ... Analysis of the loading curves revealed substantial variation in loads and maximum stiffnesses among samples from different dorsoventral regions within an individual and among homologous samples from different species. Skin thickness varied by a factor of more than five, but this only partially accounts for the differences in the force per width of sample at a given extension. Qualitative differences in the dermal collagen fibres are implied by the shapes of the loading curves and the terminal elastic modulus which varied from about 15 to 585 MN/m2. ... The size of the scales within a skin sample was not a reliable predictor of the loading behaviour of the sample. Correlations between the mechanical behaviour of skin and specializations in locomotion and associated musculature are discussed ”

[9](emphasis mine). As usual, just ask me (or at WP:REX) if anyone needs access to the full paper. My understanding is that comparing snake skin to human, with respect to ductility and elasticity (two different kinds of "stretchiness") is (somewhat) analogous to comparing different weaves or knits of the same fiber -- they can have vastly different mechanical properties, even though they are (mostly) made of the same stuff. E.g. most cotton sweaters can stretch much more than cotton T-shirts, even though they are both made of cotton. Things like gauge of the yarn, and pattern of weave, have a huge effect! SemanticMantis (talk) 00:34, 4 December 2013 (UTC)

• Keep in mind that the scales we see are not the skin itself, but overlapping (for the most part) and embedded in it. The skin itself resembles chicken skin. If a snake swallows a large meal, the scales go from overlapping to widely separated.

• Using the clothing analogy, we could design clothes that could expand that much, using elastic, pleats, etc. StuRat (talk) 10:23, 4 December 2013 (UTC)