Below, we respond to the queries by Prof. Fritz Vollrath:
------------------------------------------------------------------------

Reviewer #1 (Reviewer Name):

Fritz Vollrath

Interesting, well written and argued paper, potentially important if/when
the arguments are supported independently. I would suggest revising points where cause and effect are confused. Moreover, the discussion might be shortened to reflect the results. Alternatively the many purportedly
supporting ideas might be phrased as hypotheses that can (and should) be
independently tested.

specifics
properties of a normal pendulum and, in an ideal case, requires no muscle
to move the 71 body center of masses (BCM) steadily forward [9]. Mechanical power of upside-down 72
Q: really: NO muscle is required??

A: This is the ideal case. However, we have now written “little muscle” instead of “no muscle” as to be on the safe side.

(Fig. 1). Therefore, since body mass does not constrain as much the
evolution of leg 75 traits, selection can act on leg traits, such as stride frequency (leg diameter) and stride 76 length (leg length). Thus, if animals have evolved following the physics of 77
Q: I don't see how leg diameter determines stride frequency

A: Here we are taking the likely assumption of a correlation between leg diameter and muscle diameter. The relationship between muscle diameter and stride frequency is based on Hill (1938; see also Josephson 1993) model in which he shows how muscle contraction speed is an inverse function of muscle tension (which depends on muscle cross section). Thus, if diameter increases, muscle tension goes down and muscle speed goes up. A relationship between cross-sectional muscle area and maximum muscle power output (and thus maximum stride frequency) has been documented in at least horses (Moss et al. 1997). In reality, what happens here is that an increase in muscle diameter involves that the number of fibbers is increased. Here is a simple way of visualizing this. If the net mass that legs have to move is approximately constant (that is, if thicker legs contribute only slightly to the increase of the whole body mass), then for the same body mass and more leg diameter, we have higher acceleration during leg movement (pulling or pushing the leg within one stride) because we can use more fibbers to create larger forces: F=ma. An easy demonstration: take a pencil and shake it fast. Now take something heavy and try to move it so fast. You can't, because muscles are under a heavy load. The relative load for a spider with thicker legs is smaller and can thus move the legs faster. Obviously this applies only to very tiny differences in leg diameter (or once body size has been controlled for). An elephant will always move its legs at a much smaller frequency than any spider.
However, we acknowledge that the extension of the legs in spiders is also based on hydraulics and that perhaps a thinner leg could be extended faster from the pressure of liquid pumped inside it. However, this would still cause selection on leg diameter (no matter in which direction).
Despite all these arguments, leg diameter has not evolved independently of body size in either spider group (hangers vs. standers). Thus, we do not feel necessary to include all this elementary information in the ms. However, we would be willing to do it if the editor and reviewer feel it would improve the ms.

References
Hill 1938, Proc. R. Soc. London Ser. B, 126, 136
Josephson (1993) Annu. Rev. Physiol. 55, 527
Moss et al. (1997) Eur J Appl Physiol 75:193-199


mechanics explains the adaptive evolution of spider morphology. Indeed,
our results 123 suggest that leg length has been directly favoured by natural selection, since larger 124 spiders that hang from their webs have disproportionately longer forelegs relative to 125 smaller spiders; i.e., positive allometry and this effect is significantly stronger in these 126
Q: I don't see how your results have shown that natural selection has
directly favoured leg length, although you have shown some correlation

A: Allometric relationships can arise for several reasons. One of the interpretations of allometric relationships is that positive allometry between two traits suggests that one trait evolves faster than the other because it has been favoured by natural selection (i.e. adaptive explanation). Of course, other explanations are possible. However, we believe that our interpretation is correct because 1) it is predicted by a biomechanical model of pendulum mechanics and 2) we have shown that this occurs once phylogenetic relationships have been controlled for, thus suggesting an adaptive explanation.

"posture mode x body size" interaction, F1,101 = 0.91; P = 0.608). Thus,
consistent 134 with the mechanics of pendulum motion, both standing and hanging spiders have 135 evolved disproportionately longer legs, and hanging spiders have done so in a higher 136 degree. 137
Q: disproportionately to WHAT? And: what was driving the postulated
increase in leg length in the 'standers' as opposed to the 'hangers' for which you are making such a strong case.

A: We meant “disproportionately to body size” and have now added this to the text. However, the increase is postulated to be stronger in ‘hangers’ relative to ‘standers’ and no vice-versa and the hypothesis is outlined in Fig. 1 and in the text. Standing spiders have a constraint on how long their legs can be because otherwise they do not have enough force to lift their bodies efficiently. This is why large hanging spiders are so slow in our trials. We have now added “, as predicted by the constraints imposed on standing spiders (Fig. 1b).” to make it clearer.