An old "in the hand" feature for IDing Pallid from Common Swift in the hand is the extent of the tail fork. Casement, M.B Ibis 105, 266-8, 1965
Taking the ratio of T5 to T4 on Swifts photographed from underneath, I've managed to get a reliable differentiator on photographs, using Daniel Occiato's gallery of Pallid Swifts in Italy, and a combination of Tarsiger, Netfugl and Surfbirds photos of Common Swifts.
The requirements are being able to see all 5 TFs on at least one side of the tail and the TF not being violently splayed though if I can find a way to measure chords that requirement would drop. Currently I am excluding photos where the feather shaft is curved.
I've used a single bird - the Crosby pallid as a control - there were 3 occasions where I could measure the tail fork from photos, though I would have excluded one from my main data set on lack of clarity. This does give an idea of the levels of error to expect in the measurements. I've deliberately measured bird in a wide range of orientations, potentially increasing the errors due to masking of bases by under tail coverts, foreshortening etc, since in the end if the technique only works on perfectly square on photos, its no more use to ID single birds than a shotgun and an a ruler.
I need to put in a lot more measurements and want to add eastern swifts, but I'm really quite pleased with it.
I had one massive outlier in the Common Swifts, so far it was actually in the middle of the Pallids.
http://www.netfugl.dk/pictures/birds_uploaded/8270_Mursejler_Vejlerne_20051022_OK_IQ2B4813komp7c.jpg
Its a late October Danish bird
Taking the ratio of T5 to T4 on Swifts photographed from underneath, I've managed to get a reliable differentiator on photographs, using Daniel Occiato's gallery of Pallid Swifts in Italy, and a combination of Tarsiger, Netfugl and Surfbirds photos of Common Swifts.
The requirements are being able to see all 5 TFs on at least one side of the tail and the TF not being violently splayed though if I can find a way to measure chords that requirement would drop. Currently I am excluding photos where the feather shaft is curved.
I've used a single bird - the Crosby pallid as a control - there were 3 occasions where I could measure the tail fork from photos, though I would have excluded one from my main data set on lack of clarity. This does give an idea of the levels of error to expect in the measurements. I've deliberately measured bird in a wide range of orientations, potentially increasing the errors due to masking of bases by under tail coverts, foreshortening etc, since in the end if the technique only works on perfectly square on photos, its no more use to ID single birds than a shotgun and an a ruler.
I need to put in a lot more measurements and want to add eastern swifts, but I'm really quite pleased with it.
I had one massive outlier in the Common Swifts, so far it was actually in the middle of the Pallids.
http://www.netfugl.dk/pictures/birds_uploaded/8270_Mursejler_Vejlerne_20051022_OK_IQ2B4813komp7c.jpg
Its a late October Danish bird
- its one of the 1st things I did! Sadly there isn't a photo I can measure (Showing T% and T4 unbent), but its got a Common style tail with a very long outer TF (possibly)
