The voiced uvular preceded by *m- evolved to MC ng-（注：Baxter 中古拼音疑母）, like *N.ɢ (section 4.4.1.3). This is shown by xiéshēng contacts between MC ng- < *m.ɢ and MC y-（注：Baxter 中古拼音以母） of uvular origin: for example, the character 牙 yá < MC ngae（注：疑母麻韻二等） ‘tooth’is phonetic in the early graph for 與 yǔ < yoX（注：以母語韻三等） ‘give; for; and’, which we reconstruct as *m-q(r)aʔ, with a uvular initial, because of its graphic (and probably etymological) connection to 舉 *C.q(r)aʔ > kjoX（注：見母語韻三等） > jǔ ‘lift, raise’. Examples of early character forms are given in (511) (from GG2.573, 3.230, 9.692, 6.349)（注：‘GG’ 指古文字詁林）.(511)（古文字詁林字形略，後同）牙 *m-ɢˁ<r>a > *ŋˁra > ngae > yá ‘tooth’（注：‘<r>’ 指這個 ‘r’ 是中綴）與 *m-q(r)aʔ > yoX > yǔ ‘give; for; and’, pMǐn *ɣ- （注：‘pMǐn’ 指 Baxter 所構擬的原始閩語）舉 *C-q(r)aʔ > k(r)aʔ > kjoX > jǔ ‘lift, raise’邪 *sə.ɢA > *s.ɢA > *zA > zjae > xié ‘awry’（注：大寫的 A 指 Baxter 不確定它與小寫的 a 是否有區別）; also used in a place name:琅邪 *[r]ˁaŋ-ɢ(r)A > lang-yae > Lángyá (name of a mountain in Shāndōn) （注：Baxter 使用方括號表示「裡面的東西權且這麼寫」）If the Old Chinese initial of 牙 yá < MC ngae ‘tooth’ had been a velar nasal *ŋˁ-, yá would not have been an acceptable phonetic for 與 *m-q(r)aʔ > yoX > yǔ; thus we reconstruct 牙 *m-ɢˁ<r>a with the body-part *m- prefix. Better than a velar nasal, a uvular stop 牙 *m-ɢˁ<r>a also explains why 邪 *sə.ɢA > zjae > xié ‘awry’ has initial z-（注：Baxter 中古拼音邪母） in Middle Chinese: as we will show (section 4.4.3.4), *s-ŋ- evolves to MC s-（注：Baxter 中古拼音心母）; only before voiced obstruents does Old Chinese preinitial *s evolve to MC z- (section 4.4.3.3).Forms related to 牙 *m-ɢˁ<r>a > *ŋˁra > ngae > yá ‘tooth’ occur with initial [ŋ] in southeast Asian languages: for example, Proto-Tai *ŋa (Li 1977:204), VN ngà [ŋɑ B2], and Bahnar ngəla, all meaning ‘tusk, ivory’ (Norman 1988:19). This could be taken to support a velar nasal initial in the Chinese word, but we suspect these are ivory trade words borrow from Chinese after the change of *m-ɢˁ- to *ŋˁ-, with semantic narrowing, rather than the other way around.
W. H. Baxter:
The voiced uvular preceded by *m- evolved to MC ng-（注：Baxter 中古拼音疑母）, like *N.ɢ (section 4.4.1.3). This is shown by xiéshēng contacts between MC ng- < *m.ɢ and MC y-（注：Baxter 中古拼音以母） of uvular origin: for example, the character 牙 yá < MC ngae（注：疑母麻韻二等） ‘tooth’is phonetic in the early graph for 與 yǔ < yoX（注：以母語韻三等） ‘give; for; and’, which we reconstruct as *m-q(r)aʔ, with a uvular initial, because of its graphic (and probably etymological) connection to 舉 *C.q(r)aʔ > kjoX（注：見母語韻三等） > jǔ ‘lift, raise’. Examples of early character forms are given in (511) (from GG2.573, 3.230, 9.692, 6.349)（注：‘GG’ 指古文字詁林）.(511)（古文字詁林字形略，後同）牙 *m-ɢˁ<r>a > *ŋˁra > ngae > yá ‘tooth’（注：‘<r>’ 指這個 ‘r’ 是中綴）與 *m-q(r)aʔ > yoX > yǔ ‘give; for; and’, pMǐn *ɣ- （注：‘pMǐn’ 指 Baxter 所構擬的原始閩語）舉 *C-q(r)aʔ > k(r)aʔ > kjoX > jǔ ‘lift, raise’邪 *sə.ɢA > *s.ɢA > *zA > zjae > xié ‘awry’（注：大寫的 A 指 Baxter 不確定它與小寫的 a 是否有區別）; also used in a place name:琅邪 *[r]ˁaŋ-ɢ(r)A > lang-yae > Lángyá (name of a mountain in Shāndōn) （注：Baxter 使用方括號表示「裡面的東西權且這麼寫」）If the Old Chinese initial of 牙 yá < MC ngae ‘tooth’ had been a velar nasal *ŋˁ-, yá would not have been an acceptable phonetic for 與 *m-q(r)aʔ > yoX > yǔ; thus we reconstruct 牙 *m-ɢˁ<r>a with the body-part *m- prefix. Better than a velar nasal, a uvular stop 牙 *m-ɢˁ<r>a also explains why 邪 *sə.ɢA > zjae > xié ‘awry’ has initial z-（注：Baxter 中古拼音邪母） in Middle Chinese: as we will show (section 4.4.3.4), *s-ŋ- evolves to MC s-（注：Baxter 中古拼音心母）; only before voiced obstruents does Old Chinese preinitial *s evolve to MC z- (section 4.4.3.3).Forms related to 牙 *m-ɢˁ<r>a > *ŋˁra > ngae > yá ‘tooth’ occur with initial [ŋ] in southeast Asian languages: for example, Proto-Tai *ŋa (Li 1977:204), VN ngà [ŋɑ B2], and Bahnar ngəla, all meaning ‘tusk, ivory’ (Norman 1988:19). This could be taken to support a velar nasal initial in the Chinese word, but we suspect these are ivory trade words borrow from Chinese after the change of *m-ɢˁ- to *ŋˁ-, with semantic narrowing, rather than the other way around.我：OC *r 系與 *l 系在音位上互補且偶爾有諧聲關係，而二者也都經常與影母及中古軟顎組聲母諧聲。在缺乏大量資料支撐的情況下，我傾向於保守地將其視為廣義的 ‘kl-’型聲母。
牙 OC~MCC *ŋɹˁa → LCC *ŋɹˁa˞ → PCC *ŋʶʴa˞ → EMC *ɴʴa˞ → SMC /ɴʴa/ → LMC /ŋɨ̯a/；
與 OC~MCC *laʔ → LCC *ʎɑʔ → PCC *ʝɯ̯ʌʔ → EMC *jɯʌˀ → SMC /jɯɤˀ/ → LMC /iɤ/；
舉 OC~MCC *klaʔ → LCC *kɑʔ → PCC *kɯ̯ʌʔ → EMC *kɯʌˀ → SMC /kɯɤˀ/ → LMC /kɨɤ/；
邪 OC~MCC *l̃a → LCC *ʫa → PCC *zi̯a → EMC *ziæ → SMC /ziæ/ → LMC /zi̯æ/；
琅邪 OC~CC *ɹˁaŋla → LCC *ɾˁɑŋ.ʎa → PCC *ɾʶɑɔŋ.ʝi̯a → EMC *ɾʶɑɔŋ.jiæ → SMC /ɾɑɔŋ.jiæ/ → LMC /ɾɑɔŋ.i̯æ/。
本民科符號說明：
OC 指諧聲時期，大致為西周，ECC 指以「國風」為代表的春秋時期，MCC 指亂作一團的戰國時期，LCC 指秦到西漢（對音等材料表明中古「雲以邪」母在西漢表現得很有意思），PCC 指 ~100 AD（韻部開始鬆動），EMC 指 ~350 AD（母音整齊地高低四分、前後三分），SMC 指 ~550 AD 也就是切韻音系，LMC 指 ~800 AD 也就是韻圖音系。
*l̃ 指一個鼻化的 *l，說實在的我也不確定它與 *l 是否真的有區別。另外這裡的 *l 都有 [ɫ] 的傾向，舌位較為後縮，近音成分較大，爆破感不強，與 С. А. Старостин 的 *l 諧聲系、W. H. Baxter的 *ɢ 諧聲系實際上是相通的，只是寫法不同而已。