Response to Drs Foltz and DeLong's Letter to the Editor:

We thank Drs Foltz and DeLong for their comments bringing to our attention the discrepancy between the data presented in the abstract and our conclusions,³ and we regret any confusion this may have caused in the minds of readers. To clarify the ambiguity, we include here a table explicitly relating the housing densities used in our study to those recommended in the Guide.²

Body weight	Recommended no. of mice per pen*	Relative increase in housing density for 4, 6 or 8 mice per pen3
		4	6	8
> 25 g	3	33%	100%	166%
15-25 g	4	0%	50%	100%
10-15 g	6		0%	33%
< 10 g	8			0%

^*Number of mice of different weight categories per pen in duplex cages of 51.7 in.² in accordance with the recommendations of the Guide.

The data for the largest animals (that is, those > 25 g) showed that mice housed at 4 or 6 per pen demonstrated no significant differences for most of the parameters measured. In fact, none of the female C57BL/6J mice attained a weight > 25 g during the study, so they remained within the Guide recommendations throughout. No significant differences were noted between these female C57BL/6J mice and other mice housed at 4 or 6 per pen. Therefore, we concluded that the housing density of mice could be increased by up to 100%, to 4 or 6 per pen, withno apparent ill effect on wellbeing as reflected in the wide range of tests employed in our study. This did not appear to be the case for animals housed at 8 per pen.

Although we suggested, as part of our conclusions, that the housing density could be increased beyond that recommended in the Guide, at no point did we advocate for such an increase. In fact, on p 751 in the article, we cautioned against ascribing too much biologic significance to those parameters manifesting statistical significance.

This caveat is reinforced by the fact that very few parameters displayed a linear response to altered housing density. Such variability makes it difficult to suggest a unifying response to increased housing density; however, certain parameters are highly suggestive of significant responses to increased housing density and are worthy of inclusion in future confirmatory experiments.³

Furthermore, the most detrimental response to increased housing density was in those mice housed at 8 per pen. Some negative responses were seen relatively early on in the study before the mice exceeded 25 g in weight. For example, the growth curves presented in Figure 2 on p 746 suggest that group size may have had a greater role than floor area by itself.³ The question of group size versus cage size has been raised by several other studies referred to in our paper but none seems to have clearly distinguished between these two environmental factors.

Our study was preliminary, as our primary aim was to assess a range of test parameters to determine which were most helpful in detecting differences in the response of mice to altered housing densities. Indeed, we were able to conclude that some tests were useful and others were not. In addition, our data did not indicate that housing density increases were uniformly detrimental to the animals, and we stand by this conclusion. At the same time we strongly agree with the correspondents’ calls for more extensive behavioral testing in future studies and are fully cognizant of the ethical obligations and emotional (public perception) aspects of the welfare of animals used for research. These considerations, together with the study limitations that we pointed out (that is, sample size, lack of assessment of the impact of environmental enrichment, and testing of only animals housed in single-sex groups) further induced us to recommend caution in interpretation of our results. However, we consider the data interesting and believe that our cautiously drawn conclusions are reasonable and provide food for thought about mouse wellbeing and how to assess it.

Finally we take issue with the correspondents’ reference to Foltz and colleagues,¹ in which previous studies from The Jackson Laboratory (TJL) were criticized for some of the assumptions made and resulting conclusions reached. First, none of the authors in the current study were involved in the previous studies although most of us have worked at TJL. Second, we did not feel that overall the comments¹ were directly relevant to nor contributed to the assessment of our data because we included many more physiological and behavioral tests than the earlier TJL studies. We did not include all the behavioral tests suggested¹ because as with many researchers, we had to make decisions about, and restrict ourselves to, those tests we felt we could best perform to gain reliable data within our limited budget. Comprehensive housing studies that incorporate factors such as strain differences, effect of environmental enrichment, more extensive in- and out-of-cage behavioral testing, as suggested by the correspondents, are necessary, highly desirable,and expensive to conduct. Securing funding for such complex, large-scale studies is extremely difficult, because despite the importance of this area of research, with its potential for wide-ranging impact on much other research, those funding agencies with adequate funds appear to focus their support elsewhere.

In conclusion, we agree in large part with the suggestions of Foltz and DeLong but contend that the differences in our interpretation of the data stem from differences in emphasis. Our emphasis was to identify reliable and valid parameters for evaluation of responses to changes in housing density for use in future studies. Despite our findings suggesting that increased housing density may have no or limited apparent detrimental impact on the 2 strains of mice studied, we did not advocate implementing such an increase. We are well aware of the need for further studies before concluding that one way or the other is preferable or even adequate. Our intentions were to contribute to this important and ongoing debate concerning laboratory mouse welfare.

Letters to the Editor

Letters discuss material published in JAALAS in the previous 3 issues. They can be submitted through email (journals@aalas.org) or by regular mail (9190 Crestwyn Hills Dr, Memphis, TN 38125). Letters are not necessarily acknowledged upon receipt nor are the authors necessarily consulted before publication. Whether published in full or part, letters are subject to editing for clarity and space. The authors of the cited article will generally be given an opportunity to respond in the same issue in which the letter is published.

Respectfully yours,

Anthony Nicholson, BVSc, PhD, DACVA
Nicholson Anesthesia and Comparative Phenotyping Consultancy
Rachel D Malcolm, MS
The Jackson Laboratory
Chadi Touma, Dr rer nat, Dipl Biol
Department of Psychoneuroendocrinology, Max Planck Institute of Psychiatry, Munich, Germany
Rupert Palme, DVM, PhD
Department of Biomedical Sciences/Biochemistry, University of Veterinary Medicine, Vienna, Austria
Michael V Wiles, PhD
Senior Director, Technology Evaluation and Development, The Jackson Laboratory