Gemara

(Mishnah): One may designate produce, and declare some of it to be Terumah when he wants. He may assume that the fruit is intact.

R. Elazar says, if he finds that it was lost, he must be concerned for 24 hours (that it was already lost, and his tithing was invalid).

(R. Elazar (ben Pedas)): Chachamim argue with R. Elazar.

(Mishnah): If a Mikveh was measured and found to be deficient, any food that touched people or Kelim that immersed in the Mikveh is retroactively Tamei.

This is whether it is in a Reshus ha'Rabim or Reshus ha'Yachid.

Question: It is clear from the Mishnah that Chachamim argue! Why did R. Elazar need to teach this?

Answer: One might have thought that Chachamim meant that the foods are retroactively Tamei for 24 hours. R. Elazar teaches that this is not so.

Mishnah (Mikva'os 2:1): If one was Tamei and we are unsure whether or not he immersed, he is Tamei mi'Safek. The same applies if we know that he immersed but we are unsure whether the Mikveh had 40 Sa'im, or if there were a complete Mikveh and a deficient Mikveh and we are unsure in which he immersed.

Mishnah (2): If a Mikveh was measured and found to be deficient, whether in a Reshus ha'Rabim or Reshus ha'Yachid, any Taharos (food) that relied on the Mikveh (it touched people or Kelim that immersed in it) are retroactively Tamei. This refers to (immersion from) severe Tum'ah. If it was light Tum'ah, e.g. he ate Tamei food, or entered Mayim She'uvim... he is Tahor mi'Safek;

R. Yosi is Metamei. He says that anything Tamei is in its Chazakah until we know that it became Tahor. However, a Safek about whether something became Tamei or was Metamei something, is Tahor.

Rishonim

The Rif and Rosh (3:11) bring R. Elazar ben Pedas' teaching.

Ran (DH Gemara): We are Metaher Safek Tum'ah in Reshus ha'Rabim if perhaps it did not touch and never became Tamei. Is something was Tamei, we are not Metaher it mi'Safek.

Rambam (Hilchos Mikva'os 10:6): If we are unsure whether or not a Tamei immersed, or he immersed but perhaps the Mikveh lacked 40 Sa'im, or if there were two Mikva'os and perhaps he immersed in the deficient one, he is Tamei mi'Safek. We leave someone Tamei in his Chazakah until we know that he immersed properly. Similarly, if a Mikveh was measured and found to be deficient, in a Reshus ha'Rabim or Reshus ha'Yachid, Taharos relying on it are retroactively Tamei, unless we know that it was complete at some time. This refers to immersion from severe Tum'ah. If it was from light Tum'ah, e.g. he ate Tamei food... since this Tum'ah has no source in Torah, he is Tahor mi'Safek. This is even if we are unsure whether he immersed, whether the Mikveh had 40 Sa'im, and if there two Mikva'os and perhaps he immersed in the deficient one.

Ra'avad: If he was an Av ha'Tum'ah mid'Rabanan, he is Tamei mi'Safek.

Kesef Mishneh and Mishneh l'Melech: A Mishnah (Taharos 4:11) explicitly says so. The Rambam (Avos ha'Tum'ah 15:1) agrees. Here he was Metaher only Vlados Tum'ah mid'Rabanan.

Mishneh l'Melech: Why do R. Shimshon and the Bartenura say that we are stringent (only) about an Av ha'Tum'ah that is Metamei him mid'Oraisa? It seems that they mean that we are stringent only about an Av ha'Tum'ah, even mid'Rabanan, for Avos ha'Tum'ah are Metamei people mid'Oraisa.

Poskim

Shulchan Aruch (YD 201:65): If a Mikveh is Muchzak to become deficient at times (and one immersed in it), she must immerse again unless we know that it was complete. However, if it is not established to go below 40 Sa'im at times, even if the amount of water fluctuates, she need not immerse again. However, it is proper to check before immersing.

Source (Ran Teshuvah 66, brought in Beis Yosef DH Kosav ha'Ran): Women immersed in a Mikveh Muchzak to have 40 Sa'im. The next day it was found to have exactly 40 Sa'im. Surely, they need not immerse again. We leave the Mikveh in its Chazakah (Gitin 31b). A Chacham forbade because the amount of water in the Mikveh often fluctuates. Perhaps it had less than 40 Sa'im at the time. She is Muchzekes Temei'ah until we are sure that she immersed. Indeed, if the Mikveh is Muchzak to become deficient at times, its Chazakah is prone to change, and she must immerse again. We learn from Kidushin 79b. If it only fluctuates, but it is not established to go below 40 Sa'im at times, we leave the Mikveh in its Chazakah, for nothing suggests that it was deficient. However, it is proper to verify before immersing. We do not rely on Chazakos when the matter can be verified (Pesachim 4a. Regarding Bedikas Chametz, If possible we ask the landlord, and do not rely on Chazakos.)

Taz (85): This connotes that we must know that the Mikveh initially had 40 Sa'im. It seems that it suffices if we know now that it now has 40 Sa'im. Perhaps the Ran merely discusses the case he was asked about.

Magen Avraham (OC 8:11): The Rosh says that the custom is to rely on Chazakah that one's Tzitzis are Kosher. Zealous people inspect them before blessing on them. The Bach says that this is because there is no 'nature' that they are Kosher. Really, it is because we do not rely on Chazakah when one can check, like we say about a Mikveh. Further, even if one has no time to check, since strings are prone to snap, there is no Chazakah.

Gra (8:17): Indeed, we do not rely on Chazakah when one can check. Mikveh is not a proper proof, for there, there is Chezkas Tum'ah of the person.

Shulchan Aruch (71): If we are unsure whether or not a Tamei immersed, or he immersed but perhaps the Mikveh lacked 40 Sa'im, or if there were two Mikva'os and perhaps he immersed in the deficient one, he is Tamei mi'Safek. We leave a Tamei in his Chazakah until we know that he immersed properly.

Taz (86): We are always stringent about a Safek mid'Oraisa. Here we are stringent even if it is mid'Rabanan, for we leave Tum'ah in its Chazakah.

Beis Yosef (DH v'Chosav) the Rambam rules like R. Yosi.

Rebuttal (Mishneh l'Melech Hilchos Mikva'os 10:6): This is a textual error. The Rambam explicitly rules like the first Tana! Also the Taz erred, for the Shulchan Aruch discusses only Tum'ah mid'Oraisa.

Shulchan Aruch (ibid.): Similarly, if a Mikveh was measured and found to be deficient, in a Reshus ha'Rabim or Reshus ha'Yachid, Taharos relying on it are retroactively Tamei, unless we know that it was complete at some time.

Shach (145): Normally, we are lenient about Safek Tum'ah in Reshus ha'Rabim. Here we are stringent, for we leave Safek Tum'ah in its Chazakah.