Commentary for Avodah Zarah 79:16

The mishnah allows a Jew to buy a tarit fish which has not been minced from a non-Jew and eat it. The Talmud explains that the fish’s head and backbone must be recognizable. This is how we can be sure that a non-kosher fish did not get in there. The mishnah allowed “brine containing fish” as long as that fish was a specific fish, called the kalbit fish. This fish is a sign that the brine is kosher. In open barrels two of these fishes are necessary, but in closed barrels, one is sufficient.

R. Huna and R. Nahman argue over whether both the head and backbone need to be recognizable or whether one is sufficient. But R. Ukba argues that the kashrut of fish is determined by their having fins and scales, not by their heads or backbones. Abaye answers that while, of course, kosher fish need to have fins and scales, the fish can usually be told apart by their heads. If their heads are the same as unclean fish, then one will need to examine the fins and scales. But if we know the head belongs to a kosher fish, then the fish can be eaten.

The rules regarding the brine are more lenient than the rules regarding eating the fish itself. To eat the fish itself, one needs to be able to see the head and backbone.

R. Zera seems to hold like R. Huna, according to whom the head and backbone must be recognizable. But he used to think that the disagreement was over eating the fish itself. R. Huna would have agreed that when it comes to dipping in the brine, only one sign of its kashrut is needed. But then when he heard that they disagree over dipping in the brine, and that R. Huna said that even then both signs are needed, he stopped dipping in the brine.

R. Papa had said that the head and spine of each and every fish must be recognizable. But the baraita seems to imply that even if only one in an entire barrel is recognizable, they are all kosher.

R. Papa said that the case of the barrel of fish which Rabban Shimon b. Gamaliel permitted was only when the pieces fit together such that we could see that they all came from one fish. But this is a bit too obvious to even need to be stated. The Talmud solves this minor problem by suggesting that we might have been concerned that this was just happenstance, that there was some unclean fish together with the clean fish. Therefore, according to R. Papa, we learn that we are not concerned about such a possibility. But if all of the pieces do not fit together, then we need to recognize the head and backbone of each and every fish.

A boat-load of fish called “tzahanta” comes in to town, but it is unclear whether these fish are kosher. There are no identifiable signs of kashrut on the fish themselves (probably because they are minced), but they can see scales on the sides of the boat. R. Huna b. Hinnena declares the fish to be kosher, assuming that the scales on the sides of the boat come from the fish in the boat. But Rava notes that many fish have scales. How do we know that the scales come from these fish and not from others? Each amora issues declarations prohibiting or permitting the fish. What a conflict!

According to R. Papi, R. Huna allowed only the brine. He did not allow the eating of the fish itself. But R. Papa said that R. Huna was more lenient—even the fish could be eaten. R. Ashi is torn. On the one hand, he respects what R. Papa says, so he cannot rule strictly. Only the other hand, he has a tradition attributed to Ulla according to which the fish itself cannot be eaten unless the head and backbone of each fish is recognizable. So R. Ashi is left paralyzed, and cannot issue a directive one way or the other.

The kalbit fish is a sign that the brine is kosher. So if the barrels are open, one kalbit fish is sufficient for all of the barrels to be presumed kosher. But if the barrel is closed, then only that barrel can be considered kosher. The others must be suspected as coming from non-kosher fish. Interestingly, R. Hinena calls Rav, Shmuel and R. Yohanan “verses” as if to say that they are as reliable as the written word.

“Roe,” is another word for what we usually call caviar—the fish eggs. Since we cannot tell whether they came from a kosher or non-kosher fish, one should buy them only from an expert, one who is reliable to sell kosher fish. The same is true for entrails, which we discussed earlier.

The issue here is whether unclean fish have “roe,” meaning lay eggs. According to Rav, they do. Otherwise, roe could only come from kosher fish. But there is a baraita that says that unclean fish do not lay eggs.
There are two responses to this difficulty. The first is to emend the statement of Rav and remove the word “roe.”
The second is to suggest that they both lay eggs but that they lay them differently. Kosher fish lay eggs externally. Non-kosher fish have eggs but they hatch inside their bodies. The technical word for this “ovoviviparous” (cool word!). This is true of sharks and rays and a few other species of fish, all of which are not kosher.

Why does roe have to be bought from an expert? Aren’t there signs that distinguish kosher eggs from non-kosher ones?

If the roe is intact, it can be identified by its shape whether it is kosher or not. But if it has been pressed into a mass, the shape is lost and it should only be bought from someone who is reliable.

R. Dostai held that non-kosher fish do not lay eggs and therefore the word should be erased from Rav’s statement. But the baraita above which notes the different shapes between kosher and non-kosher roe assumes that non-kosher fish do lay eggs.
R. Dostai would answer by correcting the baraita in a different way. So if we’re going to correct it, rather than correcting it to read “fish eggs” correct it to read “fish entrails.” Now the baraita explains how to identify kosher fish entrails. But are fish entrails round and pointed? The answer is that the bladder of a kosher fish does indeed have this shape.