The Vedanta Sutras of Badarayana, Commentary by Sankara (SBE38), tr. by George Thibaut  at sacred-texts.com
6. (Should it be said that the souls are not enveloped by water) on account of this not being stated by scripture, we refute the objection on the ground of those who perform ishtis, &c., being understood.
Well, let it be granted that, on account of question and answer, water, passing through the forms of sraddhâ, &c., may in the fifth oblation obtain the shape of man. But still we cannot allow that the souls when moving from one body into another are enveloped by water. For this is not directly stated by scripture, there being in the whole passage no word referring to the souls, while there are words referring to water. Hence the assertion that the soul goes enveloped by water is unfounded.--This objection is invalid, we reply, 'on account of those who perform ishtis, &c., being understood.' For in the passage beginning 'But they who living in a village practise sacrifices, works of public utility and alms, they go to the smoke' (V, 3, 10), it is said that those who perform ishtis reach, on the road of the fathers leading through smoke. &c.. the moon, 'From ether they go to the moon; that is Soma, the king.' Now these same persons are meant in the passage about the five fires also, as we conclude from the equality of scriptural statement in the passage, 'In that fire the devas offer
[paragraph continues] sraddhâ. From that oblation rises Soma the king 1.' To those 2 (persons who have performed ishtis, &c.) water is supplied in the shape of the materials employed to perform the agnihotra darsapûrnamâsa and other sacrifices, viz. sour milk, milk, &c., which substances, as consisting mostly of water, may directly be considered as water. Of these, when offered in the âhavanîya, the subtle parts assume the form of an apûrva resulting from the oblation 3, and attach themselves to the performer of the sacrifice. Then (when the sacrificer dies) the priests offer his body, with the funeral ceremonies 4, into the crematory fire, with the mantra, '(may) he (go) to the heavenly world, svâhâ.' Then the water forming the oblation--which has connected with deeds resulting from faith 5--having assumed the form of an apûrva envelops the souls of those who had performed the sacrifices, and leads them up to the heavenly world to receive their reward.--In accordance with the preceding interpretation scripture says in the agnihotra chapter also--in the complementary passage constituting the reply to the six questions--that the two agnihotra-obligations go up to the other world in order to originate the fruit (of the work of the sacrificer), 'Those two oblations when offered go up, &c.' (Sat. Br. XI, 6, 2, 6).--Hence we conclude that the
souls, when going to the enjoyment of the fruits of their works, are enveloped by the water of which the oblations consist 1.
But how can it be maintained that those who perform sacrifices, &c., go to the enjoyment of the fruit of their works, considering that scripture declares them when having reached the moon--by the path leading through smoke, &c.--to become food, 'That is Soma the king; that is the food of the gods; the gods do eat it' (Kh. Up. V, 10, 4); and the corresponding passage, 'Having reached the moon they become food, and then the Devas feed on them there as sacrificers feed on Soma as it increases and decreases' (Bri. Up. VI, 2, 16)? If, in accordance with these passages, they are eaten by the gods as by tigers, &c., it is not possible that they should enjoy the fruit of their deeds.--To this the following Sûtra replies.
109:1 Both passages speak of something reaching, i.e. becoming the moon. Now, as that something is, in the passage about the road of the fathers, the gîvas of those who have performed ishtis, &c., we conclude that by the sraddhâ also, from which in the other passage the moon is said to rise, those gîvas are meant, or, properly speaking, the subtle body of water which envelops those gîvas.--Dhûmâdivâkye pañkâgnivâkye ka somaragatvaprâptisravanâviseshâd ishtâdikârinah sraddhâsabditâdbhir veshthitâ dyulokam yântîti bhâtîty arthah. Ân. Gi.
109:2 An. Gi. introduces this clause by: nanu mahad iha srutyor vailakshanyam, sraddhâsabditânâm apâm kvakid dyuloke homah srutah kvakid ishtâdikârinâm dhûmâdikramenâkâsaprâptir na ka teshâm âpah santi yena tadveshthitânâm gatis tatrâha teshâm keti.
109:3 I read, with a MS. of An. Gi., âhutyapûrvarûpâh.
109:4 The so-called antyeshti.
109:5 And is on that account properly called sraddhâ.
110:1 Saṅkara's attempts to render plausible the interpretation of sraddhâ by 'water,' and to base thereon the doctrine of the souls when going to a new body being enveloped by a subtle involucrum of water (and the other elements contained therein) are, of course, altogether artificial. I do not, however, see that he can be taxed with inconsistency (as he is by Deussen, p. 408). Sraddhâ is to him in the first place the gross water which constitutes the chief material employed in the sacrifices; in the second place the apûrva which results from the sacrifice, and which is imagined to consist of the subtle parts of the water whose gross parts have been consumed by the sacrificial fire. These subtle parts attach themselves to the soul, accompany it as an involucrum when it goes to another world, and form the base of any new body which the soul may have to assume in accordance with its previous deeds.