In simple terms, phone is any sound that can be made. It can be a word or noise or animal cry. Conversely, phoneme is an individual speech unit and is related to a particular language. For example, EIGHT - phone is /eit/ ; phonemes are /e/, /i/ and /t/ It's easy to understand if you think of sounds that animals make. Take a dog, for exammple. The phone it makes is the same in all languages, but it is written differently from one language to another and different phonemes are used.

By convention, animal sounds, etc. do NOT count as "phones", because "phone" comes from a Greek word "phoné" that meant human voice. A "phone" is a minimal SEGMENT of speech sound: the smallest chunk of simultaneously occurring speech noises that can be "cut out" from the speech stream. For example, to produce an [m] several organs must cooperate, with their peculiar 'noises' - you must expel air, your vocal cords must vibrate, you must close your oral cavity, the air must pass through your nasal cavity, you must stop it with your lips, and you must let it come out through your nostrils. The [m] can, therefore, be analysed as a continuant, voiced, bilabial, nasal sound - in opposition to [p], [b] or [n], which share some of those features but lack others. Think of all those "speech features" sounding simultaneously and imagine a cut at the point where they stop; that is a "segment", and a "phone". ["segment" and "speech feature" are different concepts]. A phone will vary depending on what its neighbouring phones are - for example, in English, [m] is different in "small", "immediate", "lamb", "meet", "rims", etc., and you may yourself pronounce [m] differently even in the same word on successive occasions, but such differences are accepted as yielding mere 'tokens' of the SAME phone. A "phoneme", in contrast, is the MINIMAL SEGMENT of speech THAT CANNOT BE REPLACED BY ANOTHER WITHOUT CHANGING MEANING. E.g., in English, /m/ and /n/, or /n/ and /ng/, similar enough as they are, are different phonemes, because "thin" and "thing", or "neat" and "meat" have different meanings. So, your criterion: two "phones" point to different "phonemes" if exchanging them LEADS TO A CHANGE OF MEANING. Whereas the number of phones is infinite, the number of phonemes is fixed, and small - between twenty and fifty, depending on the language. Each language has its own 'system' of phonemes, all ‘in opposition to’ each other depending on the "distinctive speech features" (voice, nasality, labiality, etc.) they have/lack. If you compare languages, the ‘types’ of phones used in each will not completely coincide, certain differences between phones will suffice to change meaning (i.e., become ‘phonemic’) in one, but not the other, etc.