Wednesday, October 13, 2010

Picky pipefish dads and the problem with correlations

A pipefish paper made a big splash when it first came out, but how convincing is the evidence? Here is my take.

Love is a battlefield. For some animals, the battle goes on even after love has been made.

I'm talking about post-copulatory sexual selection. Sexual selection after mating is a lot like sexual selection before mating, except whereas before mating, males compete with other males and females choose which male to mate with, after mating it is the sperm that compete with other sperm and the females reproductive tract that choose which sperm to use to fertilize eggs. The later phenomenon is known as cryptic female choice because you can't see it happening unless you do a paternity analysis on the offspring [1].

The stereotypical competitive male and choosy female are products of differences in the potential reproductive rate for males and females, and are not necessarily a result of having testes or ovaries. Sex-role reversed species such as syngnathids illustrate this point quite nicely.

This pregnant male seahorse is a syngnathid.

In syngnathids, the males "get pregnant": the females lays her eggs into the male's pouch and he carries them around until they hatch. This male pregnancy lowers the male potential reproductive rate below that of females. As a result, it is the females that compete with each other for access to males, and it is the males who are more choosy about who they mate with [2].

A recent paper published in Nature by Pazcolt & Jones [3] claims that in one syngnathid, the Gulf pipefish, males practice cryptic male choice after sex by favoring eggs from "attractive" females and aborting eggs from "unattractive" females.

The Nature News coverage of this article was titled "Male pipefish abort embryos of ugly mothers" which at first might makes you wonder how such subjective concepts like "ugly" and "beautiful" were assigned to pipefish by human observers. But actually it is just a catchy headline, in the original paper, attractiveness simply equals female size, based on the fact that male pipefish are known to prefer large females.

When it first came out, the paper was covered in several blogs (including by Ed Yong and grrlscientist). However, since then the papers conclusions have been challenged by some big names in the field of sex-role reversal. Gywnne et al. [4] claimed that the results of Pazcolt & Jones could be explained entirely by larger females having higher quality eggs. Pazcolt & Jones replied [5], defending their original conclusions.

Let's dive into the experiment and see what conclusions we can make for ourselves:

The experiment consisted of taking male pipefish through two rounds of mating, getting pregnant, and giving birth. They varied female size during the two matings such that males mated with either large or small females.

They then measured a bunch of things about the second brood and correlated them all to each other. So many correlations.

Figure 1. Reprinted by permission from Macmillan Publishers Ltd: Nature (Paczolt, K. A. & Jones, A. G. (2010). Post-copulatory sexual selection and sexual conflict in the evolution of male pregnancy. Nature 464, 401–404.), copyright 2010.

From these figures and other (ungraphed) data, we can see that offspring survival was positively correlated with:

  • Female size
  • Number of total eggs laid by female (which correlates with female size)
  • How quickly the male was willing to mate with female (also correlates with female size)

The authors conclude that males chose to support offspring from larger females that they liked better (and/or abort offspring from smaller females that they didn't like). But I think (and Gywnne et al. agree with me) that the most obvious conclusion for this result is that larger females lay more eggs and their eggs are more likely to survive, which is probably why males prefer to mate with larger females in the first place. Both conclusions are supported by the results.

The authors actually address the alternative conclusion, for some reason NOT in the main article but instead in the Supplementary Info. They state:

"The reduced survivorship of eggs originating from small females could be explained if small females produce eggs with exceptionally low viability."

But the authors exclude this explanation based on the following:

  1. All of their females were above average in size compared the the field population from which they were collected.
  2. The huge discrepancy in average offspring survival between the large and small female groups: 50% for the small females, and 89% for the large females.
Hmm, seems convincing. Let's check the methods and see how different the two classes of females were in size:

Small females = 9.3-10.6 cm
Large females = 10.8-12.2 cm

Is there really a difference between a 10.6 cm pipefish and a 10.8 cm pipefish? I am skeptical.

Another reason I am skeptical is that the 50% survival for the small females is an average. If you look at the actual data (Figure 1D) you can see that the small female average is being driven down by 3 females whose eggs ALL DIED. Those three broods might have been cryptic male abortion in action, or else there could have been something seriously wrong with those eggs. But this experiment unfortunately can't distinguish between the two.

There was another thing that Pazcolt & Jones found that I haven't mentioned yet. They found a negative correlation between prior brood survival and current brood survival, in other words that the more offspring that survived in one pregnancy, the less offspring survived in the next pregancy and vice versa. Basically this indicates that being pregnant takes energy. This makes sense because male pipefish transfer nutrients to the developing baby pipefish inside of the pouch. After one pregnancy, the male will have energy-debt if he dedicated a lot of energy to offspring survival, or will have an energy-surplus if he didn't. Pazcolt & Jones hypothesize that males could choose to reduce investment in one brood (such as from a smaller female) in order to save energy for the next brood.

Of course it is not clear whether the male is "choosing" to dedicate more energy to certain brood or whether broods require different amounts of energy depending on their size. Consider a single dad raising triplets versus a single dad raising only one baby. The dad will triplets will probably be spending way more energy on parental investment by necessity (because there are three of them!). Or consider a pregnant woman carrying triplets versus one carrying only one child. Which woman would you expect to eat more ice cream???

Once again, the results could be explained just as easily by large females producing more and healthier eggs. Since larger females have larger broods, then the male could end up allocating more energy to those broods _just because they are larger_, irrespective of what they thought of the female.

Before we can make a definite conclusion on whether cryptic male choice occurs in pipefish (and I am guessing that it probably does), we need more experiments that disentangle the effects of female egg quality from male differential investment. I don't think it would be that hard to design such an experiment, it could be as simple as testing the same female with a smaller and a larger male. The small male should find the female very attractive while the larger male should not. This would control for differences in female egg quality because the same female would be used in both matings.

And if you have any clever ideas for an experiment, please do comment ;)

^[1] Ben-Ari, E.T. (2000). Choosy females: Exploring the role of cryptic female choice in sexual selection and battles over paternity. BioScience 50, 7-12.

^[2] Eens, M., and Pinxten, R. (2002) Sex-role reversal in vertebrates: behavioral and endocrinological accounts. Behav. Process. 51:135-147

^[3] Paczolt, K. A. & Jones, A. G. (2010). Post-copulatory sexual selection and sexual conflict in the evolution of male pregnancy. Nature 464, 401–404

^[4] Gwynne, D. T., Judge, K. A. & Kelly, C. D. (2010). Evidence for male allocation in pipefish? Nature 466

^[5] Paczolt, K. A. & Jones, A. G. (2010). Evidence for male allocation in pipefish? Pazcolt & Jones reply Nature 466


  1. It's interesting that there was a negative correlation between the survival of a first brood and a second brood, I wonder if there may be a maximum brood size that a male is willing to take on at any given time, from say an aberrantly large female, as overall fitness may be affected if the cost is going to be too high in future broods.

    You man enjoy this post on Dr Bondar's blog about male abandonment in Nicaraguan cichlids.


  2. Thanks for the link, Rhaco, I always enjoy reading about cichlids and not just cuz I study them :P