Gemara

160b (R. Yochanan): If words were Teluyos (written between the lines), and they are Mekuyam (the document mentions them) at the end, it is valid;

An erasure invalidates a Get, even if it is Mekuyam at the end. This is only if it is in a place fitting and large enough for it to say 'Sharir v'Kayam.'

161b (R. Yitzchak bar Yosef): An erasure (even if it does not invalidate a Get, due to its place or small size) must be Mekuyam in the Get.

167a: A document said 'a third b'Pardisa (of an orchard); the bearer erased the top and bottom of the 'Beis', leaving a 'Vov', so it would say v'Pardisa (and an orchard).

Abaye: Why is there space between the 'Vov' and 'Pardisa'?!

Abaye forced him, and he admitted to his forgery.

Rishonim

The Rif brings our Gemara.

Nimukei Yosef (DH Omar ha'Mechaber): The Yerushalmi says that we attribute everything possible to an erasure. This is why it mentions here only erasures, for this includes what was scraped. There was not such a need to mention Teluyos, for if one was was not Mekayem them, the document is Kosher, just we do not learn from the Teluyah. If an erasure or scraping was not Mekuyam, the entire document is Pasul, lest there was a Tanai there. Tosfos, the Ritva and Ro'oh say so.

Rambam (Hilchos Malveh 27:8): One must be Mekayem every erasure at the end of the document, and say 'this letter or word or line is written over an erasure or between the lines, and it is all valid.' If an erasure is in a place for and large enough for it to say 'Sharir v'Kayam', even if it is Mekuyam, it is Pasul. Perhaps he erased it, added text, and was Mekayem it with space between the text and the witnesses.

Magid Mishneh (citing the Rashba): Some of my Rebbeyim say that we did not say 'any erasure or Teluyah invalidates a Get', for a Teluyah that was not Mekuyam does not disqualify the document, just we do not learn from it. An erasure that is not Mekuyam invalidates a Get, for perhaps a Tanai was written there and it was erased. I disagree. Also an erasure that is not Mekuyam does not disqualify the document, just we do not learn from it. A Tosefta (11:5) says that 'an erasure or Teluyah invalidates a Get only if it is mi'Gufo.' Our Gemara mentioned erasures, but the same applies to Teluyos.

Beis Yosef (CM 44 DH v'Im): The Rashba (1035) says that 'mi'Gufo' is the name of the lender, the borrower, and the sum.

Magid Mishneh: It seems that also the Rambam equates erasures and Teluyos. A place for 'Sharir v'Kayam' is the next to last line. The Rambam connotes that a document does not need Sharir v'Kayam. It is only if there are erasures and Teluyos, to enable learning from them. Also, he holds that one should not write the Kiyum of erasures and Teluyos in the last line, for we do not learn from it. This is clear from the Gemara.

Rosh (10:2): It seems that there is a blank line under the end of the text (above the signatures), it is Kosher even if the last line does not recap the document. When there is no space in between it must recap, for we could not learn from the last line. If there is a blank line in between, we are not concerned lest the witnesses signed two lines below and the bearer added a line. Gitin 19b asked 'why is a Persian document Kosher? The last line must recap the document!' We answered that it does. We could have answered that there is a space between the text and the witnesses. However, it seems that if there is no recap it is Pasul even without concern for forgery. We asked similarly (Gitin 11a) about a Persian document given in front of Yisrael witnesses, according to R. Elazar, who says that Edei Mesirah Karsei (witnesses who see a document given empower it). There is no concern for forgery, for if the bearer adds something, the witnesses will (remember what was written and) realize this, even after many days! Likewise, even if there is a blank space under the text and the witnesses, the last line must recap, for we decree due to a document without a line in between. Chachamim enacted to recap, without distinction. The Rashbam says that we disqualify a document with two lines between the text and the witnesses, even if witnesses say that nothing was added, for the document is unlike Chachamim enacted. The Sugya proves that any document that could be used to swindle is Pasul, even if no fraud was done, for it is unlike Chachamim's enactment. Therefore, if witnesses sign on one line, they may not leave a space to the side (before their signatures), lest the bearer cut off the document, and write what he wants to the side. A document with the text and witnesses all in one line is Kosher! Rather, there must be no space next to the witnesses, or they should dirty it with ink. We are concerned lest witnesses signed on an ink spot (and not on the document) only when it is between the text and the signatures, like the concern lest witnesses signed on a greeting under the document (176a). Or, witnesses may sign one under the other (even if there is space to the side). Even though the bearer could write 'I owe 100 to you' to the side, and (it will look as if) one witness signed on it, we are concerned only when the bearer can forge a document that can be used to force others to pay. We are not concerned if he can force the witness to pay, for the witness caused his own loss. However, if one may join two documents, each of which was signed by one witness, we must disqualify a document with space next to the witnesses.

Rashba (2:254 DH Od d'Amrinan): An erasure between the lines or letters does not disqualify a document due to concern lest a receipt was written there. We learn from the man who changed a Beis to a Veis. Abaye did not disqualify the document due to the erasure, until he forced him to admit. One could reject the proof by saying that the erasure was not evident; this is weak. 5.

Poskim

Shulchan Aruch (CM 44:5): If there are erasures or Teluyos or words that the scribe wrote over, one must be Mekayem them before Sharir v'Kayam.

Shach (13): The Kiyum must be before Sharir v'Kayam only if it is in the last line.

Rema: One need not write 'this is their Kiyum'.

Shulchan Aruch (ibid): If one was not Mekayem them, the document is Kosher, but we do not learn from them.

Beis Yosef (DH v'Chosav bi'Teshuvah): The Rashba (Toldos Adam 56) says that if one erased, scraped and wrote the same letters, he did not disqualify the document, but he acted like a schemer. In Teshuvah 100 (attributed to the Ramban) he says that a scraping between lines or on the back does not discredit a document. However, one should threaten the bearer (into admitting what was there), for receipts are often written between the lines, and there is supporting evidence (that this is why he scraped).

Gra (11): Another proof is from 167a. A man changed a Beis to a Veis, to make it say v'Pardisa; Abaye did not disqualify the document (until he admitted, like the Rashba above - Birkas Eliyahu).

Rema: Some say that if an erasure was not Mekuyam, the entire document is Pasul, lest the bearer erased something that weakens him. If it is clear from the document what was erased, the document is Kosher.

Shach (14): The Nimukei Yosef says so in the name of the Yerushalmi and many Gedolim. This is primary. If not, we must be concerned lest one erase an unfavorable clause or Tanai. The Rashba learns from the Tosefta that it is Kosher. I say that this is only when it is clear from the document what was erased. Even so, if it was mi'Gufo, it is Pasul. If it is not clear from the document what was erased, and it is possible that there was a Chiyuv or unfavorable clause or Tanai, this is like mi'Gufo, and it is Pasul. Anything that the entire document depends on is mi'Gufo. Based on this, we can resolve a contradiction in the Tur and Rema. Here they disqualify the entire document if an erasure was not Mekuyam. In EH (125:19) they bring only the Rambam's opinion, that if it is in the Tofes, the Get is Kosher. We must say that this is when we can infer was what erased. The Magid Mishneh says that the Rambam equates erasures and Teluyos. I say that he equates them only to obligate Kiyum (but only erasures disqualify the document).

Shulchan Aruch (ibid): If they are in Gufo of the document, i.e. the name of the lender, borrower and the sum, the document is Pasul if they were not Mekuyam. If they are in the date and they were not Mekuyam, it is like a document without a date.

Tur (45:30): The Rosh (68:17) says that since we often write a receipt between lines, one must ensure that there is no scraping between the lines. We must be concerned lest payment was written there, and he erased it. All the more so, if payment is between the lines on erased parchment, we must be concerned that more was written and he erased it, and wrote what he wanted. If witnesses say that the borrower himself wrote the payment, there is no concern.

Darchei Moshe (44:2**): The Rivash (382, brought in Beis Yosef Reish Siman 44) says that if the quill was passed again over a word, it is as if it is Teluyah, and we do not learn from it, and the rest of the document is Kosher. 'It is like a Teluyah' implies that an erasure disqualifies.