This is what a friend of mine said:

When there's human offspring, it will look a rough 'medium' between the two parents, with it sharing features from the two. well if the offspring of one set of parents has sex with the offspring of another set, the outcome becomes a 'medium' of those two. So the range from what we shall call 'completely average looks' (a hypothetical state of the average of everyone in the world, which is acceptable as we're looking at a process of millions of years) becomes less and less the further into the future we go. Eventually, there is almost no range between people and 'completely average looks'.

In summary: take the hypothetical average of everyone in the world, and because the offspring of two people is a slight mixture of those two people, they are a tiny bit closer to the hypothetical average of everyone in the world, which means that eventually people will be getting closer and closer to this average, so will one day look the same. So he's suggesting that slowly people will become more and more similar.

My argument is that, although this baby may be that tiny bit closer to the overall average, they're new look has effectively altered the average, meaning that they aren't actually getting closer to the average.

I'd like an opinion and flaw from some experts, such as yourselves, so please...

  • $\begingroup$ What he's getting at is called homogeneity, for future reference. At least, I think so. 'Completely average looks' means very little in a biological context since evolution works on a genetic scale. $\endgroup$ – MCM Feb 22 '13 at 0:28
  • 1
    $\begingroup$ It sounds like they are saying all brothers would look alike, but that isn't true either. $\endgroup$ – shigeta Feb 22 '13 at 7:02
  • $\begingroup$ Edited for what he actually meant, cos i misinterpreted it $\endgroup$ – ODP Feb 22 '13 at 13:56

Quantitative genetics is one reason why this doesn't happen. For example height is affected by many genetic loci, not just 1 gene, and we can for the sake of the following illustration call them Locus A through Locus J - 10 Loci.

Now this is purely illustrative and hypothetical - in actuality there are some differences between this example and the real world, but nevertheless I shall plough on. Imagine everyone carries a basic set of genes which makes them 150cm tall. Loci A-J each have two alleles, either T or t, with one inherited from each parent. Each T allele adds 2.5cm to the basic height per copy (each parent has two copies, one goes to the offspring) and t has no effect. The parents are both heterozygous at each loci, that is they have one T and one t at each pair of loci. Therefore both parents are 175cm tall.

When these parents mate the offspring inherit one copy of each parent at random. As such the most likely single outcome is that the offspring will inherit 10 T's and 10 t's - roughly five of each from each parent. However, purely by chance they could receive 10 T alleles from each parent, giving them a height of 150+(20*2.5) = 200cm. Likewise, a sibling could also inherit only t alleles thus giving them a height of 150+(0*2.5)= 150cm. Thus variation can actually increase from one generation to the next.

There are other reasons why we do not regress to the mean as well, but I am writing a quick answer over breakfast, but I will be in the office soon where I can get some more thorough explanations from text books. E.g. (Dis)assortative mating, Mutation, Recombination, Environmental variance, Selection. However, with the increase in global movement it has become easier for dominant alleles to spread which is why the number of blue eyed people in the US is decreasing.

Here is an interesting article which backs up what I said about quantitative genetics, called "Will humans eventually all look like brazilians?" Also for good quantitative genetics texts take this as a basic intro and this for further reading.

  • $\begingroup$ Also, I think we many of the answers to this question are the same as would answer the question "why don't siblings look identical?" - simplified to "polymorphism exists." $\endgroup$ – rg255 Feb 24 '13 at 23:02
  • $\begingroup$ Please look at my edit, which is in bold, as well as the new paragraph below it. Do you have anything more to say? $\endgroup$ – ODP Feb 24 '13 at 23:24
  • 2
    $\begingroup$ well for everyone to look the same it would require that they are a) genetically identical, this would require ubiquitous inbreeding (everyone homozygous for the same allele at all loci) and b) everyone experiences the same environment all the way from development Hutterites but you will still see "considerable" variation within those - basically it is impossible for us all humans to be identical because we have so many people so inbreeding is low, mutation occurs in the DNA regularly, and complete inbreeding would likely cause the species to collapse at the hands of disease. $\endgroup$ – rg255 Feb 25 '13 at 8:47
  • 1
    $\begingroup$ @ollyprice and, if you read my original answer you will see that even if both parents are average (175cm) it does not necessarily mean that their offspring will be average (anywhere from 150-200cm, and possibly further with mutations and environmental variance). $\endgroup$ – rg255 Feb 25 '13 at 8:48

The basic point your friend is missing is that offspring appearance is not simply a mixture of their parents'. Each child also has individual characteristics which were not present in the parents. This is due to a variety of processes, the most important of which is chromosomal cross over.

Each offspring will be a mixture of its parents and its own individual characteristics. Therefore, your friend's argument does not stand because each generation causes novel characteristics to appear so each successive generation changes in unpredictable ways.

  • $\begingroup$ Please look at my edit, which is in bold, as well as the new paragraph below it. Do you have anything more to say? $\endgroup$ – ODP Feb 24 '13 at 23:25
  • 1
    $\begingroup$ @OllyPrice your friend's entire argument is based on a wrong assumption. Offspring are not simply the average of their parents. Consider siblings, you can have a couple with 18 children, each of these will be different not simply the average of their parents. Mutation and cross over and simple errors in DNA replication ensure this. $\endgroup$ – terdon Feb 25 '13 at 18:17

Here's another take on it:

Suppose, for starters that you have only one gene that is responsible for your trait. Say, having version 1 (denoted T) gives you +1 to height, and having version 2 (denoted t) gives you -1. Suppose further that each version occurs in half the population. Now every individual has 2 copies of each gene, so that 1/4 of people has TT for +2, 1/4 has tt for -2 and 1/2 has Tt or tT for 0 change in height.

Now suppose two people with Tt or tT mate. If traits followed simple averaging pattern your friend suggested we would get the average of 0 and 0, that is 0 - all the time. But in genetic model (which is what really happens) 1/4 of the ofspring will get TT for +2, 1/4 will get tt for -2 and only 1/2 will actually get tT ot Tt for 0. Note that the variance - amount of variation - increases (everyone was 0 before mating, but some of the offspring has +2 or -2), whereas with the "mixing" model this does not happen.

For another mating paring - TT with tT - 1/2 of offspring will have TT for +2 and 1/2 will have tT for 0, effectively replacing the parents, so that the amount of variation stays the same (since the mean of the population is 0 variance contribution of these two individuals is 2^2+0^2=4 in both parental and offspring populations). Note that the mixing model would have all offspring at +1, which decreases the amount of variation (variance contribution going from 4 to 1^2+1^2=2).

In general, the mixing model of mating always decreases variance - leading to the "everyone looks average" result in the long run, but the actual genetic mating keeps variance constant (unless another factor - like preferential mating, mutation, or environmental factors intervenes; this was pointed out in other answers).

The same thing happens if the proportion of T and t are not 50-50. This is called the Hardy-Weinberg principle.

Finally, if there are many genes that contribute to height, the variance in a trait (like height) can be modeled as the sum of variances from each gene. Now if each one of the genes is inherited through the same genetic mechanism that preserves variance, the total variance will also be preserved. So we will not actually all become "average".

  • 1
    $\begingroup$ You should add to that the fact that height does not only come from genetic traits, the environment (nutrition, pollutants, stress, etc) plays an enormous role there. $\endgroup$ – nico Mar 23 '13 at 6:34
  • $\begingroup$ That is most certainly true. I suppose the point of the above is that the part of the variance that comes from genetic factors does not decrease. What happens to the variance due to environmental factors is a completely separate matter, of course. $\endgroup$ – Max M Apr 11 '13 at 4:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.