Relevant

Epiphanius in the 300s criticized the Gospel according to the Hebrews (GATHM) for replacing locusts with honey-cakes (manna) in what today is identified as Matthew 3:4: “Thus they were resolved to pervert the truth into a lie and put a cake in the place of locusts.” (Epiphanius, Panarion 30.13.4-5.)

It is humorous he thought this implied heresy. But Epiphanius’ anger is our blessing because the solution to this variance proves GATHM was written in Hebrew and came before our Greek translation. For Epiphanius insists the Gospel according to the Hebrews does not say locusts, as does the Greek Matthew, but said cake instead.

What word was Epiphanius reading in GATHM that made him translate it as cake? The Hebrew word Epiphanius read in GATHM was manna. This was normally translated as honey-cakes. So in Greek, for example, manna is rendered as egrkis—honey-cake—in the Greek Septuagint of 247 BC of Exodus 16:31. As Professor Tabor explains: “the Greek word for ‘honey cake’ (egrkis)...is used for ‘manna’ that the Israelites ate in the desert in the days of Moses (Exodus 16:31.)” (James D. Tabor, The Jesus Dynasty (Simon & Schuster) at 349 n. 9.) See below.

Why does this prove GATHM’s manna came first and not locusts? Because after a proper first translation egrkis—honey-cake—for manna in the Hebrew original, the next Greek copyist made a one-letter error, writing akris meaning in Greek locust. Professor Tabor explains it is clear that a Greek copyist copied egrkis (manna / cake) incorrectly as “locust”—akris—which represents a single letter scribal error. (James D. Tabor, The Jesus Dynasty (Simon & Schuster) at 349 n. 9.)

This is how a Greek translator turned manna in Hebrew into locust in Greek. The original text necessarily was the Hebrew word manna. Then it was rendered correctly into Greek as egrkis—a honey-cake. But then it was incorrectly copied as the Greek word akris (locust). Thus, manna in Hebrew became by mistake locust in the Greek-based translation we read today. It could not be the other way around—that GATHM was a Hebrew translation of a Greek text. If a Greek edition came first, and then a Hebrew translation was made, then egrkis (locust in Greek) would be arbe in Hebrew. The word arbe would never end up down the line mistakenly being spelled as manna in a later Hebrew copy. Hence, GATHM clearly came before the Greek version of Matthew, and was the source of our modern Greek version of Matthew.

2020 Addendum: The Same Goes for the Greek Mark.

This same conclusion destroys the familiar Marcan priority claim. For the Greek Mark at 1:6 equally has "locusts." Whether this is from the Greek Matthew, or vice-versa, is now irrelevant. What it proves is whichever provided the source for the other, the Hebrew Matthew was the superior original source to both Greek Mark and Greek Matthew.