In linguistics, the morphophone is the basic unit of meaning smaller than a word, accounting for all the ways that unit is pronounced across all dialects of a language. It is to a morpheme what a diaphoneme (or diaphone) is to a phoneme.
That's a mouthful, and this is better understood with examples than definitions.
In English, the letter "t" can be pronounced a bunch of different ways, and those pronunciations usually differ according to the speaker's dialect. "Button" can have a glottal stop in the middle, "bu'on," in a Cockney accent, but in a Mid-Atlantic accent it will have a crisp, sharp /t/. "Water" may have a /t/, a /d/, or an alveolar tap in the middle, among other options. The /t/ can be aspirated or not. It can be fronted closer to the teeth, a dental /t/ more than an alveolar /t/. An English speaker will still recognise all of these sounds as contained by the concept "the letter T," however, and won't compartmentalise them as having different meanings. This totality of "T-ness" is the diaphoneme "t."
Now let's try this with a morphophone, expanding the "-ness" concept beyond a single consonant. We'll use the word "right," since it only contains one morpheme. In general, no matter how an accent conveys the sound of "right," a native English speaker from any part of the world will still understand it as "right," whether it sounds like /ɹaɪt/ in Michigan, /ɹojt/ in Australia, /rait/ in the westernmost reaches of Wales, or /ɹɑt/ in Texas and Oklahoma. Think about how "right" can sound. "Rite," "royt," a trilling "rrrrite," a drawling "rawt" or "rot." Whether the word itself is "write," "right," or "rite" makes no difference, since all of these dialects consider those words to be homophones to one another.
Occasionally, dialects will disagree about which words are homophones, resulting in the pin, pen and merry, Mary, marry differences between regions on the same landmass. This presents linguists with a bit of a challenge, because it casts doubt on the universality of morphophones, and by extension, diaphonemes, within a given language... and this further calls into question the precise definition of how "dialect" and "language" differ from each other, and where the lines are drawn between dialects and languages with high mutual intelligibility but low diaphonic relatedness. As with most things in linguistics, this is all very abstract and prone to extreme variation, with no "right" answers to apply to all circumstances.
Iron Noder 2017, 15/30