Two variants are listed, ειπαν and και ειπαν, and in addition there is plenty of support for the text. The problem is that the apparatus gives the impression that the orthography ειπαν is widely supported (unlike ειπον in, for example, 8:52). However, of all the witnesses listed in support for the text and the two variants  only D has ειπαν, with all the others reading ειπον. 

Arguably it is difficult to reflect orthographic variation (and I don't have a simple solution), but in this particular instance the suggestion given by the apparatus is unusually misleading.

Dirk,
Thank you for your input.  Since the third edition, when Weymouth was replaced with Weiß, the Nestle text has ειπαν probably because of the agreement of Westcott-Hort and Weiß.  The first two editions of the Nestle agree with Tischendorf and Weymouth in reading ειπον.  But you're right that the text should be changed to ειπον since only D has ειπαν.  And since you raise a good point about orthographic consistency, the matter of orthography will perhaps be an agenda item for the next NA/UBSGNT editorial committee meeting.