I'm working on creating a species that favours small families for a number of reasons, and have no interest in 'breeding as much as possible'. As a consequence, their population size is small. But how small can a population be, without running (ever) into problems with genetic diversity? I'm interested in more than 'survivability'; the species' gene pool should not degenerate (some margin allowed). I'm looking for some well-founded answers or directions.

Assume that genetics for this species works similar to that of humans, and that their gene pool is of good quality at the start. You can also assume that they choose partners wisely, but there is no designation of partners, so margin is necessary. They have good health care and low child mortality rates. Assume that at least 90% of the population reaches at least fertile age.

Other 'rules':

  • Average number of children born is two (might be increased to three if that is necessary to sustain the population)
  • Species is monogamous
  • No genetic manipulation

This question differs from the question What is the minimum human population necessary for a sustainable colony? because:

  • It is assumed (for a population of 80 / 160) that the population will mingle with a larger population after ten generations
  • The population of 350 still resulted in some genetic issues (of which some disastrous)

This question differs from the question How many humans do you need to maintain a population indefinitely? because of the given context that everyone in the group is the genetic parent of exactly two children each with a different partner which is violating above assumptions.

This question differs from the question Minimum Population For a High Tech Society? because it is not about genetic diversity, but the skills and intelligence needed to run a high tech society (which is possibly a follow-up question).

Some other questions touch this topic, but they are (slightly) different in either the assumptions or the direction of the question.

  • $\begingroup$ would this be of some use as a reference newscientist.com/article/… $\endgroup$
    – P Chapman
    Nov 22, 2017 at 20:27
  • $\begingroup$ That article assumes that the population comes into contact with a larger populations after ten generations or so. So not entirely what I'm looking for. $\endgroup$
    – Century
    Nov 22, 2017 at 20:51
  • 1
    $\begingroup$ One of the key issues is how consciously managed the mating process is with regard to genetic diversity. You can maintain genetic diversity for a long time is you bar inbreeding sufficiently, but in a small population, especially, if everyone is not in synch timing-wise, this may mean that lots of people may end up mating with someone who isn't their first choice and you also have to manage early childhood exposure of children to other children in the community so that they don't end up seeing their genetically optimal mates as siblings (i.e. the Westermarck effect). $\endgroup$
    – ohwilleke
    Nov 23, 2017 at 3:02
  • $\begingroup$ 'Creating a species' indicates it does not have to be human. A great portion of the answer depends on how many chromosomes,and how many genes? This would give an idea of the genetic complexity of the species in the first place. For example, a species with only one hair color, eye color, skin color, height range, etc. (that is, a very homogeneous species) would not require a large gene pool to maintain diversity as there is only limited diversity to begin with. In other words, every procreation would be inbreeding. $\endgroup$ Nov 25, 2017 at 1:18
  • $\begingroup$ The issue then becomes one of maintaining a sufficient population to prevent deleterious degrading mutations from becoming established. That is, sufficient genetic material to maintain healthy genes. $\endgroup$ Nov 25, 2017 at 1:19

3 Answers 3


I happen to work in population genetics with a focus on conservation. In the past I have worked in human genetics also (human populations were really small for most of the existence of our species).

The general gist of the answers above are quite right.

There is a rule in conservation genetics that says you need at the very leas an effective population size of ~50 to have a viable population. Do not confuse the technical term "Effective population size" with "Census size". Effective population size models assume a certain number of theoretical ideas, like e.g. random mating.

There are several ways you can keep genetic diversity up:

  • The least inbreeding the better. Monogamy is actually better than random mating and way much better than "winner takes all".
  • Make sure that over generations people mate with distant individuals (even if from a small population). Mating with your sibling is not a great idea to maintain diversity
  • If you allow say, an immigrant per generation than you are probably safe. There are plenty of studies that, if a migrant (a single one!) enters the gene pool and has reproductive success, that is enough to maintain genetic diversity in very small populations.
    • Your population sex ratio (number of born males over born females) should approach 1

Note that if your world can tolerate minimal migration, than you are mostly sorted.

  • $\begingroup$ Hi Tiago, and welcome to Worldbuilding. It's great to have subject matter experts around; I hope that we will see more from you in your area(s) of expertise! Genetics is a subject fraught with misunderstandings (and I definitely don't pretend to be an expert myself!), so it's great to have people around who are actually able to both spot those errors and explain how things really work. $\endgroup$
    – user
    Nov 25, 2017 at 12:11
  • $\begingroup$ Interesting points, especially about monogamy being better than random mating and that immigrants can make such a difference. I can work with this. $\endgroup$
    – Century
    Nov 28, 2017 at 19:45
  • 1
    $\begingroup$ If I recall monogamy is better if the sex ratio approaches one (it does with humans, for example. And it does with most sexual species). $\endgroup$ Nov 28, 2017 at 20:07
  • $\begingroup$ There is a generic drive for the sex ratio to approach one. $\endgroup$ Dec 7, 2017 at 21:30
  • $\begingroup$ Men can produce more children, which encourages a higher number of make births, but embassies are certain to produce children. Which leads to the 1 to 1 ratio in male to female births in most species on earth. $\endgroup$ Dec 7, 2017 at 21:31

Well, first things first: If 90% of those born reach reproductive age and have offspring of their own (which is all that matters in the long term), and each pair has two offspring, then mathematically the species will be declining. Plugging 2÷0.9 into my calculator says you'll need at least about 2.23 births per pair just to maintain a steady-state population with a 90% reproductive rate, regardless of what happens to the remaining 10%. Let's round that to 2.25 and every fourth pair will need to have one additional offspring for a total of three (as 4*0.25 = 1 and 3*2 + 1*(2+1) = 9). In this case, the most common number (mode) of offspring per pair will be 2, with an average (arithmetic mean) of 2.25. That's your absolute minimum to ensure that the population doesn't decline simply from lack of reproduction.

However, that leaves you with next to no margin at all for any unforeseen events. If anything happens, then you need more offspring elsewhere in the population that end up reproducing in order to compensate for the loss. If a mother dies in first childbirth, for example, and this causes the death of the offspring (not necessarily a given with good healthcare, but definitely a risk), then you need someone else to have four (or two to have three each) where ordinarily they would only need to have two offspring each.

I would strongly suspect, but lack citations for backing that up, that if you start out with a healthy gene pool, genetic degredation over ten generations should be sufficiently small to not cause major issues in the short term. Researching breeding in some uncommon breeds of dog, where the effective population size especially within a country can be on the same order of magnitude as that which you are discussing, may provide some insights here. (While dogs are obtained across borders for the purposes of breeding, both as adults for a single or a few litters or as pups by breeders hoping the individual will turn out well, even e.g. within the EU there are significant obstacles for this, causing such exchange of genetic material to happen less frequently than it could and consequently reducing the effective population size.) However, over longer time periods, this becomes a significant concern with a small starting population and such a low reproductive rate. At that point you need to consider what would be a minimum viable population over your desired time scale with an appropriate probability of survival, into which a potentially huge number of factors plays, but genetic degeneracy and accident rate (and accident survival rate) would probably be major factors. We already have a number of questions and answers discussing MVPs so I'm not duplicating that here. For some inspiration, while the situation obviously isn't exactly the same, you may also want to look at the wolf population on Isle Royale (see also here, here and here, among many others; use your favorite search engine for more), which is probably one of the more interesting unplanned experiments in the genetics of an isolated population.

  • $\begingroup$ Good points about the number of children, I already added that I could increase to three, but now I see your calculations, I think that mode was what was going on in my head, not average. Because average is not going to work with 90% of course. And you do need some more margin, as you state, so three is better. Can I edit my question? I do miss what you suggest as a starting population though. $\endgroup$
    – Century
    Nov 22, 2017 at 20:40
  • 1
    $\begingroup$ @Century You already stated that three is allowed if required. I'm mostly just trying to illustrate what is the minimum that you need for your population to survive, and how to get there with as little change as possible to your desired two per pair. I think that answers your question from the angle of reproductive rate required. As for the minimum number of individuals to start out with, I don't think that's all that different from simply discussing minimum viable populations in general, but again dog breeding may provide insights because many dog breeders battle such issues already. $\endgroup$
    – user
    Nov 22, 2017 at 20:53
  • $\begingroup$ That is true, and I'm happy to have some direction regarding the reproductive rate. Regarding the minimum viable populations, I haven't found something that stays within the boundaries of my assumptions. But I will look into your posted link with posts :) $\endgroup$
    – Century
    Nov 22, 2017 at 20:58
  • $\begingroup$ @Century I just plugged "minimum viable population" into the search box at the top of the page. You might possibly get some ideas doing something similar over on the Biology SE as well. $\endgroup$
    – user
    Nov 22, 2017 at 21:00

TL;DR -- Population can be arbitrarily small, but there are risks

Here's the thing...

Lack of genetic diversity is a problem, but it's not an automatic death sentence. Low g.d. means two things:

  • Your population all has the same (or very similar) disease resistance profile. That means they're more at risk of some pandemic wiping them all out.

  • Your population is similar genetically. That means an increased chance of reinforced negative recessive genes showing up, making for sickly offspring.

These can be overcome with proper animal husbandry (or heck, human husbandry!). This is in fact how new breeds are created -- heavy inbreeding, with culling of specimens who are sickly or don't fit in. Note that inbred wild species don't get the advantage of culling; instead you have sick animals running around. The culling happens, but it's a lot slower and hurts the overall population more. And in royal house populations, you can't just go around culling the Crown Prince, right?

The popualtion of wisent (European bison) was down below 50 at one point, and has been brought back up to several thousand. See: https://en.wikipedia.org/wiki/European_bison .

So you can have an almost arbitrarily small number (for sanity, let's say a few hundred), so long as a few rules are followed:

  • Keep up your medical technology as best you can. You might not get a plague, but if you do, it'll be a doozy.

  • Be prepared for some real heartbreak. Many children will be stillborn, or have some genetic disease, compared to Earth numbers. But over time, with (shuddersome, but good for fiction) sufficient euthanasia, you will expose recessives, both harmful and good.

Interesting discussion here: https://en.wikipedia.org/wiki/Inbreeding

Money quote:

Despite all its disadvantages, inbreeding can also have a variety of advantages, such as reducing the recombination load,[16] and allowing the expression of recessive advantageous phenotypes. It has been proposed that under circumstances when the advantages of inbreeding outweigh the disadvantages, preferential breeding within small groups could be promoted, potentially leading to speciation.[17]

So, to wrap it up: Inbreeding is a problem, but a manageable one which does carry risks.

Oh, last word, just for fun. Cheetahs in Africa have such low g.d. that they can almost universally accept skin grafts from one another without problem. So your population of humans will be able to donate organs to one another much more easily...

  • $\begingroup$ Hey, equal representation!! Good husbandry AND good wifery. $\endgroup$ Nov 25, 2017 at 1:23
  • $\begingroup$ @JustinThyme I get it, I really do, but the last time I gave a seminar titled "new techniques in animal wifery" ... well. Turns out the abstract was visible in google search, and... well. $\endgroup$
    – akaioi
    Nov 25, 2017 at 3:41
  • $\begingroup$ Interesting perspective; I see inbreeding happen in horses a lot to establish good genes, and opinions about that practice vary. But I didn't think of the benefits for my question. $\endgroup$
    – Century
    Nov 28, 2017 at 19:46

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .