Skip to content

The paper discussion

Details can be found at https://github.com/richelbilderbeek/Bolnick_and_Stutz_2017.

The conclusion of the paper is that extreme body sizes fare better. This can be shown in a plot that uses the absolute of the relative normalized body mass, as shown in plot versions A (from the paper) and B (reproduced from data) below.

When taking just the relative normalized body mass, this pattern breaks down, as shown in plot version C shown below.

Instead, the conclusion should be:

  • in the lake: smaller individuals thrive
  • in the stream: bigger individuals thrive.
flowchart TD
    A
    B
    C
    A --> |reproduced| B
    B --> |using non-absolute values| C
Version Figure 2 version Description
A Figure 2, in the paper As in paper
B Figure 2, reproduced Reproduced using author's code
C Figure 2, reproduced Reproduced using the real values

Note that there is another flaw in the analysis, that is not mentioned in the original paper, which is the different amounts of fish in an enclosure. The original paper describes three fish per enclosure. In such a setting, it seems reasonable to normalize the body mass of the fish within each enclosure. However, in three enclosures (L6, L25, and S24), only two fish were present. This will always result in normalized body masses of -0.707 and +0.707 (i.e. the square root of 2 and its negative), regardless of the fish had very similar of very different body masses. The effect of this is left unexplored, as we think that using absolute values of body mass is the more influential flaw in the original paper.

References

  • [Bolnick & Stutz, 2017] Bolnick, Daniel I., and William E. Stutz. "Frequency dependence limits divergent evolution by favouring rare immigrants over residents." Nature (2017).