Referee 3’s review:

* * * * * *
N.B. These are the comments made by the referee when reviewing an earlier version of this paper. Prior to publication the manuscript has been revised in light of these comments and to address other editorial requirements.
* * * * * *

Review of the original submission:

The manuscript describes a set of studies designed to show that honeybees can learn to distinguish between different numbers of objects. I have only one concern with this otherwise nice study, but it is one that makes me think that the present set of experiments are not suitable for publication in any journal.

The problem, which the authors bring up on page 8, is that the feeder was retained during the transfer tests. There was therefore no control for the possible influence of the attractant Nasunov pheromone that honeybees release at a rewarding feeder. Such controls are absolutely basic for choice experiments. Despite the authors' claims, it is entirely possible that the attractant on the feeder biased the choices towards a positive result.

If the authors repeat the experiments, they should use dedicated test chambers, both of which contain dedicated feeders. In this way, none of the feeders or chambers will have been marked by pheromone during training.

On a more positive note, the result on page 8, lines 6-8, seems interesting to me. I hope that the authors will explore it further. It deserves a proper publication in its own right.


Review of the first revised manuscript:

One month obviously would not give anyone sufficient time to repeat the experiments with odor cues suitably controlled for, so I must again strongly recommend that the manuscript be rejected for publication. This is a pity, as the experiments are otherwise very elegant. I hope that the authors find the time to repeat the critical experiments, making sure that odor cues are not available during the tests. In the present experiments, it is not possible to say that the bees were not following odor cues, and so no useful deductions can be made about any 'visual based generalization'.

I think I said in my original report that the difference between the low numbers and the high numbers is suggestive, but not sufficient. An additional little doubt has also crept in since then. Seeing that the experiments were carried out over two full field seasons, and on two continents, I can't help also wondering whether there might have been some other differences between the two sets of experiments.

The new figure 2 shows how choices change with training. The new analysis shows that the probabilities of the second choices are much like the probabilities of the first choices. As far as I can tell, the analysis points towards the hypothesis that the second choices are independent of the first choices. By the fourth block, a bee is more likely to go towards the choice with the feeder, whether it is the first or second time it is counted. I don't see this as evidence that the bees do not use odor cues. As an aside, I am not sure whether the first levels of 70% correct are recorded after 120 minutes of training, as the figure legend would suggest, or after 3 days, as the Methods section would suggest.

I went back to examine three of the studies the authors cite to justify not controlling for odor in the present experiments. I didn't find there evidence that good controls were unnecessary in the present experiments.

In Zhang et al. 1996, the question of whether bees might have been using odor cues was controlled by testing with 4 out of somewhere between 12 and 19 boxes of a maze replaced with clean boxes. The replacements made no difference to the results. Using clean test boxes was a correct and necessary control to make. But that result cannot be used to make the generalization that honeybees never use odor cues.

In the second paper I read, Zhang et al. 1999, I didn't find any mention of a control for the use of odor. So I am not convinced by the results in that paper.

In the third paper, Zhang et al. 2005, again the authors did not use clean test chambers, as they should have done to make the results totally convincing. But there was a specific test condition controlling for the use of odor. A feeder was placed behind one of two identical patterns, and approaches were counted. But it's not at all clear that the distribution of pheromone in this control case - on the feeder, boxes and in the air - would be anything similar to what would be built up during the normal training.

In summary, bees are well known to release attractant pheromone at feeders, and to approach places where there is pheromone. It is not sufficient to remove bees when you see them releasing the pheromone. It would be simple to do all the tests in a convincing manner. As it stands, a sizeable number of people will be skeptical of these results.