[the prayer is invalidated; because it is impossible to build the remainder of it upon what has passed once the interval has become long], and if the [interval has not become long, he builds upon what has passed of his prayer]. Ahmad stated this textually in the narration of a group. Al-Shafi'i held this view, [and Malik held a similar view; and regarding the length and shortness of the interval, reference is made to custom and convention]. Some of our companions said: Whenever he omits a pillar and does not catch up with it until he has performed the salutation, his prayer is invalidated. [Al-Nakha'i and al-Hasan said: Whoever forgets a prostration from a prayer, then remembers it while in the prayer, performs it whenever he remembers it, and when he finishes his prayer, he performs the two prostrations of forgetfulness]. It is reported from Makhul and Muhammad ibn Aslam al-Tusi regarding a praying person who forgets a prostration or a rak'ah, that he performs it whenever he remembers it, and performs the two prostrations of forgetfulness. It is reported from al-Awza'i regarding a man who forgot a prostration from the Dhuhr prayer, then remembered it during the 'Asr prayer, that he continues in his prayer, and when he finishes, he performs it. Our argument, that the prayer is not invalidated when the interval is short, is that if he were to omit a rak'ah or more, then remember it before the interval became long, he would perform what he omitted, and his prayer would not be invalidated by consensus. The hadith of Dhu al-Yadayn has indicated this, so if he omits a single pillar, it is more fitting that the prayer should not be invalidated, for it does not exceed the omission of a rak'ah. The proof that the prayer is invalidated by the prolongation of the interval is that he has failed to maintain continuity (muwalat), so his prayer is not valid, just as if he had remembered on a second day. And there is no limit to the length of the interval.