Amphibians and ultra high diluted thyroxine--further experiments and re-analysis of data

A model of thyroxine and metamorphosis of highland amphibians is frequently mentioned as an example of experiments on extremely diluted substances in discussions around 'homeopathy'.

The model was scrutinized by reanalysing the results of the initial researcher A and a second researcher B as well as of 5 external researchers C between 1990 and 2013. Rana temporaria larvae were taken from an alpine highland biotope. The test solution was thyroxine 10(-30) (T30x), tetra-iodo-thyronine sodium pentahydrate diluted with pure water in 26 steps of 1:10, being agitated after each step. Analogously prepared water (W30x) was used for control. Tadpoles were observed from the 2-legged to the 4-legged stage. Experiments were performed in different years, at different times of season, and their duration could vary. Frequencies of 4-legged animals, effect sizes and areas under the curves (AUCs) were calculated and regression analyses were performed to investigate possible correlations between year, season, duration etc. Experiments were in line with animal protection guidelines.

The total set of data A + B + C as well as subsets A (initial researcher, N=286+293), B (second centre, 965 + 965) and C (5 external researchers, 690 + 690) showed an effect of extremely diluted agitated thyroxine reverse to that known of molecular thyroxin, i.e. test values were below control by 11.4% for A, 9.5% for B and 7.0% for C (p<0.001 for each of the subsets). The effect size (Cohen's d) was >0.8 (large) for both A and B and 0.74 (medium) for C.

Although a perfect reproducibility was not obtained, this paradoxical phenomenon was generally consistent in different observations. Correlations were found between details of laboratory handling, as well as environment temperature, and the size of the results.