I was thinking about the following question: why do most creatures on earth reproduce asexually or bisexually, but not trisexually?
Looking on the internet, I read an interesting perspective https://www.zhihu.com/question/303528094 that attempts to answer this question via the following mathematical model, which I call the reproduction model.
A species has $m$ sexes, and $n$ different types of sexual chromosome, which we call $\{1,2,\cdots, n\}$.
Each individual of the species has a genotype of $m$ sexual chromosomes, which can be described as an $m$-element multiset with each element in $\{1,2,\cdots, n\}$. There are disjoint families $F_1, F_2, \cdots, F_m$ of genotypes, such that an individual exhibit sex $i$ iff their sexual chromosome lies in family $F_i$.
There is a distribution $\mu$ on $F_1 \sqcup \cdots \sqcup F_m$ that describes the distribution of each possible genotype across the entire species. We require equal number of individual of each sex: we must have $\mu(F_1) = \mu(F_2) = \cdots = \mu(F_m) = 1/m$.
The main point of the argument is that the distribution $\mu$ must be stable. Specifically, consider the following reproduction process: for each $i$, we sample a random individual from sex $i$ according to $\mu$. We then form an offspring by selecting a uniformly random chromosome from each individual. Then the genotype of this offspring must also be distributed according to $\mu$.
Let me give two stable examples.
- Asexual reproduction: $m = n = 1$, and $F_1 = \{(1)\}$. This is clearly stable.
- Bisexual reproduction: $m = 2, n = 2$, $F_1 = \{(1, 1)\}$, $F_2 = \{(1, 2)\}$, and $\mu$ is uniform. This is also stable, since during reproduction there is an equal chance that an offspring of type $(1,1)$ or $(1,2)$ are born.
The author of this models argue that trisexual reproduction is not stable, by considering the following model:
$m = 3, n = 2$, $F_1 = \{(1,1,1)\}, F_2 = \{(1,1,2)\}, F_3 = \{(1,2,2)\}$, $\mu$ is uniform. Then the probability that the offspring is of sex $1,2,3$ is $2/9,5/9,2/9$ respectively. For example, the offspring is of sex $1$ only if all three parents donate type $1$ chromosome, which happens w.p. $1 \cdot 2/3 \cdot 1/3 = 2/9$. So this model is unstable.
However, the author said that they cannot rule out models with more than $2$ chromosome types i.e. $n \geq 3$.
So my question is
Does there exist a stable reproduction model with more than $2$ sexes?