Here in both manuscripts, Oliver Cowdery initially wrote were; then he crossed out the were and inserted was supralinearly with the same level of ink flow (the crossout itself in both cases has heavier ink flow, but this is because Oliver crossed out the were several times in order to ensure its deletion). It seems that Oliver’s inclination here was to write the subjunctive “if the seed were”, even in both manuscripts; but in each case he caught his error virtually immediately and restored the was.
In if-clauses elsewhere in the text, we have examples of both was and were when the subject (or delayed subject) is singular. There are eight occurrences with the singular was in the original text for which the was has been maintained in the text:
On the other hand, there are at least 36 original occurrences with the subjunctive were. Some of the were examples are otherwise virtually identical with was examples:
These parallel examples show that either was or were is possible in if-clauses with singular subjects. And since Oliver Cowdery showed no predilection to emend the eight invariant instances of “if X was” to “if X were”, it seems reasonable to assume here in Alma 32:36 that his virtually immediate correction of were to was in both manuscripts was simply the result of his attempt to get the text down correctly.
Summary: Accept in Alma 32:36 Oliver Cowdery’s corrected was in the if-clause (“if the seed was good”); the correction is found in both manuscripts and appears to be virtually immediate.