I was idly scrolling through some old Facebook posts when I came across a post describing the additional 24 nest boxes that we installed at Edgecombe flats in 2021 where I was reminded that we painted every other chamber black inside. See here.
I had forgotten about our experiment, but in 2023 we surveyed these boxes and found 16 of them occupied. This was a properly designed experiment. The double boxes were numbered 1 to 12 with each chamber labeled e.g. 1L or 1R. All odd-numbered double boxes were black on the left inside and all even-numbered black on the right. This should eliminate any preference that the birds may have for left or right from confounding the result.
I checked the data and found the following results:
Black vs unpainted interiors
There were 8 pairs in the painted boxes and 8 in the unpainted ones.
The probability of getting exactly this result, if the Swifts have no preference would be 0.333, so less likely than an unequal result!
This adds to a similar experiment at Trumpington Meadows in 2019 where in 10 boxes, 5 black and 5 unpainted, there were 4 pairs in the unpainted boxes and just 1 in the painted boxes. The probability of this result is 10.3%
It does seem that there is little point in painting the inside of Swift boxes black.
More experiments like this would be useful corroboration.
Left vs right chambers
Swifts used the left chamber vs the right chamber in the ratio 9:7.
Combination comparisons
Comparing the categories 'black left', 'black right', 'unpainted left' and 'unpainted right' yields occupied boxes in the ratios 5:3:4:4.
Double vs single occupancy
The number of double boxes with pairs in both chambers was 4, so 8 pairs. The other 8 pairs occupied just one chamber. In other words, the 16 pairs had spread themselves as widely as they could, occupying all 12 double boxes. If they had no preference for having a double box to themselves, the probability of this result is 17.23%, so maybe they would prefer single boxes. [Acknowledgement to Martin Newell for help with statistics]
