Probably a trivial question but why are two options provided for output bit order in the convert fingerprint component?

I cannot see what difference this would make when calculating the similarity between the fingerprints resulting from either method
Under what circumstances would the selection of one be favoured over the other?

Thanks in advance

Kris