The forms for the second person are derived from the Arte (68). Throughout this section the accent marks are quite erratic. In several places, for example, Collado has snata and even snat.
 In the material which follows Collado has brought together items from several sections of the Arte; for the interrogatives see (65-65v), the indefinites (66), and the demonstratives (68).
 These reduplicated forms are not derived from Rodriguez' description and are apparently misstatements of the forms care and are which would otherwise be missing.
 An abbreviated form of monomsu; cf. Arte (139v).
 Collado is here speaking with reference to the normal order in Latin.
 The treatment of the verbal system by Collado follows in a general way the Arte (6v-54v). In the material that follows specific references will be made when a comparison of the two works is suggested.
 The text has secundae coniugationis. This error, which is repeated throughout the text, is not present in the Spanish manuscript.
 The text again has secundae coniugationis.
 This list covering the Kami-ichidan and Kami-nidan verbs is derived from a similarly defined sub-group of the first conjugation in the Arte (28). Since the verbs cabi, sabi, and deqi are in no way indicated as extraordinary in Rodriguez' presentation, I have amended the text to include their present tense form.
 The text reads for this gloss fucore afficior. The proper word is mucore 'mould,' with the literal translation being 'I am affected by mould.'
 The Dictionarium has this verb listed as kami-nidan, xij, uru, and therefore not exceptional.
 Cf. Arte (7) where a similar list is presented.
 For the source of Collado's description of the future tense cf. Arte (7v).
 The text reads secundae coniugationis.
 Rodriguez more correctly has this rule as the root plus i or yo; e.g., aguei or agueyo. The form aguei is used by Collado in the construction of the optative below.
 This form is correct but does not follow his rule for the formation of the imperative (see note 79).
 Rodriguez has baquemono 'evil spirit' and the Spanish manuscript baqemono, rather than banguemono 'soothsayer.'
 Extracted from Rodriguez' version of a sentence in the Amakusa edition of Esop's Fables (p. 417). The original reads, Arutoqi Xantho chinsui xite yraruru tocoroye, fitoga qite daicaino vxiuouo fitocuchino nomi tucusaruru michiga arǒcato tni,... 'One time when Xantho [Esop's master] was drunk, a man came and asked if there was a way to drink all the waters of the ocean in one swallow....' it is abbreviated by Collado in such a way as to obscure the construction.
 Also apparently extracted from the Esopo (p. 477). The original has, ... riǒbǒni tachiuacarete yru tocoroni qitunega yosocara coreuo mite, futatuno nacani vocareta fittu jiuo totte curǒta, 'when they [two lions] had gone their separate ways, the fox, seeing this from afar, took the sheep which had been between the two of them and ate it.' By changing riǒbǒ to nhb Collado created a less than satisfactory example.
 Modeled on Iyeuo idzuru tocorouo cubiuo quiri votoita 'when he went outside his head was cut off.'
 Modeled on Missauo asobasaruru tocoroye vjei faxe atumatta 'when mass was being celebrated, many came running and gathered around.'
 Apparently modelled after Arte (20v) nantomo voxiare caxi 'whatever you say,' with the imperative formation again confounded.
 Rodriguez (25v) specifies the location of this usage as Chūgoku, Bungo, Hakata, and other Ximo districts.
 This example, together with so zonze na below, reflects the loss of a distinction between z and j which was taking place during this period.
 The text has secundae coniugationis.
 The Arte (27) records here aguenedomo, aguenuto mǒxedomo, aguezutomo, aguenebatote, and agueidemo. Neither aguenaidemo nor the participle aguenaide, below, are found in the Arte, though they are attested to elsewhere. Cf. Yuzawa Kōkichirō, Edo kotoba no kenkyū (Tokyo, 1954), p. 626.
 This rule, derived from Rodriguez (Arte, 29), is misformulated by Collado. Rodriguez' rule is correct; change the nu of the negative present to i. It is formulated correctly for the third conjugation, below.
 Collado's rule clearly confuses the formulation of the present with that of the future. Significantly in the Arte Rodriguez never refers to the future forms of any verb other than his model narai. If Collado had had access to the Arte Breve he would have found (41) the following principal parts for vomoi; vomoi, vom, vomta, vomovǒ, vomoye. The only other use in the Ars Grammaticae of this form is on page 62 where Collado has the incorrect form vomov. The manuscript does not record this form.
 Although Collado's transcription permits this rule to yield the appropriate forms, it obscures the fact that the final i of the root is a vowel, while the i of the imperative is a semivowel. Rodriguez' transcription better reflects the phonological facts; naraye, vomoye, and cuye.
 This completes Collado's treatment of the third negative conjugation. The two paragraphs which follow are part of his treatment of the substantive verb. There is no section heading for the affirmative substantive verb; and clearly a portion of the text has been deleted. The Spanish manuscript (cf. Ōtsuka's 1957 edition, p. 45) includes a new section which begins by recording the following substantive verb forms; ari:aru, gozari:gozaru, i:iru, and vori:voru.
 Collado's presentation of the substantive verbs is obscure. The text reads: Verba ver substantiua sunt, gozaru, gozaranu, voru, uori nai, dea vel gia: deuanai, aru:aranu, vel, gozaranu uoru rinai, & .... The translation attempts to punctuate the list to reflect the contrast between affirmative and negative forms. The main confusion is the apparent effort to contrast voru and vorinai. Voru (glossed by the supplement of the Vocabulario as estar, and used in the Dictionarium as the gloss for existo, etc.) is not used by Rodriguez in the Arte. Vorinai (unglossed in the dictionaries) is clearly defined by Rodriguez as the negative of the polite verb voriaru, which is derived by him from von iri+aru (Arte, 165v). Possibly Collado had intended to contrast voru with voranu and voriaru with vorinai but confounded the two pairs and then repeated his error at the end of the list; or again he may, in the absence of Rodriguez' guidance, have simply misunderstood the matter. Putting the alternative forms aside, the list should read gozaru:gozaranu, vori aru:vori nai, gia:devanai, aru:aranu, and voru:voranu. Collado's treatment is patterned only loosely after the Arte (2v-6v).
 Collado seems to be unaware of the irregularity of vonaji.
 Collado is following the general rule established on p. 10 for such forms as caij. He might better have followed Rodriguez who would transcribe canaxǔte, as do we.
 The missing 'closed o' aside, Collado's transcription of this form with an n is indicative of the clarity with which he perceived the nasalization in this context.
 Cf. Arte (18v-19v).
 The text reads c vaau ni voite va, with the errata changing the verb to cuvazu.
 This historically inaccurate rule is derived from the Arte (18v).
 In the one example of this construction, on page 62, Collado has the form tovazunba.
 The original is in the soro style; Iǒjǒni voiteua uquetori mǒsubequ soro.
 Cf. Arte (19v).
 Here and throughout the section Collado transcribes as ro the potential particle which should correctly be written r (cf. Arte, 11v). It will be noticed that all but one instance of the 'open o' on p. 35 of the text has been left unmarked.
 Collado has derived this list from the Arte (45-47). His terminology is, however, rather misleading. What he classifies as verba irregularia are those which Rodriguez considers deponent, that is verbo defectiuo, with the term verbo irregular being used by Rodriguez for the adjective. Given this misunderstanding Collado begins his list with an explanation of the irregularities of qi,uru. This verb is on Rodriguez' list only because "it lacks certain forms in the affirmative" (45v). Rodriguez has a list of 43 deponent verbs, beginning with tari, from which Collado has selected the first 14 and then a few from the remainder.
 In the restricted context of an adjectival; cf. modern arayuru koto.
 Cf. Arte (45v) where Rodriguez transcribes vreyeyo.
 Loc. cit. Rodriguez presents vreru as an alternative form for vre in the present tense and then selects that variant for the infinitive.
 Formation (formatio) is to be understood here in the sense of derivation, and diversity (differentia) in the sense of class membership.
 The opening paragraphs of this section follow the Arte (68-70 and 96-108v). The list of particles, beginning with maraxi, follows 160-168.
 The text, here and in the next sentence, reads secundae coningationis.
 The form dojucu is incorrect. It is taken by Ōtsuka to be dōshuku 'a person living in the same house.' The Vocabulario records the item djucu 'a young boy who serves a priest.' Djucu best fits Collado's translation.
 The text again reads secundae coniugationis.
 Cf. Arte (160-164) from which this list and the following material have been derived.
 Throughout his treatment of the respect language Collado glosses his verb forms in the first person, even though that translation might be inappropriate to any context.
 Rodriguez (Arte, 162v) specifies the distribution of vo and go, using gosacu atta as his example of the construction in context of a Chinese vocabulary item. Collado does not refer to this distinction.
 The text reads secundae coniugationis.
 The text reads secundae coniugationis.
 Ōtsuka (1957) suggests maraxi is correct and alters the example. Since the list begins with maraxi, I assume the error to be in the citation.
 The material for this section is derived from the Arte (164v-168).
 While the material for this section is drawn from various sections of the Arte, the bulk of the particles and their descriptions are derived from Rodriguez' treatment of postpositional (73-77) and adverbial constructions (112v-125).
 Rodriguez' list (77v) runs as follows; vchi, voi, faxe, ai, tori, mexi, tui, and voxi. On the basis of Collado's examples voxi should have been included in his list.
 Collado's transcription qinpen is phonemically correct while being phonetically less accurate than Rodriguez' quimpen.
 Collado has altered Rodriguez' version from Nippon, even though the Dictionarium glosses consuetudo japonica as Nippon catagui.
 Collado, in the Dictionarium and here, prefers mmu to uma.
 This particle is not described in the Arte.
 Rodriguez (Arte, 116) records Core coso yocarǒzure and states that in this context coso has the same meaning as Queccu and Cayette.
 Cf. the Arte (117) where the list is given as Reba, Ni, Tomo, the potential, and Te.
 Rodriguez' version runs Iesu Christo fitono vontocoroua. (For Collado's use of reduplicatiuus see note 53.)
 As the first example indicates, the zzu variant is not restricted to the negative preterit, but is the form which appears for da in all contexts, as here with the preterit of iomu.
 In the absence of other examples it is not possible to determine if Collado assumed the present tense form to be iuru or uru. The correction here follows the spelling used consistently in the Arte.
 Both Collado and Rodriguez agree that verbs ending in tai govern the accusative case; cf. Nanigaxiuo yobitai (Arte, 14v).
 The text reads secunda persona.
 Rodriguez has Vatacuxiua nantomo buchfde tofǒ ga gozanai [... buchfǒde ...].
 Rodriguez uses the transcription gorǒjerarei in the example from which this sentence is derived. (The ten other occurrences in the Arte have goran.) The Dictionarium uses only goron, while the Vocabulario lists both goran and goron. The Spanish manuscript has goran.
 Rodriguez has mairade canauanu.
 The Arte has the plain form mairǒcotode attaredomo.
 The Arte has mairumajiqueredomo.
 Perhaps an attempt to follow the rule, established in the syntax below, that states the v of the accusative particle is lost after n. If this is the intent, the comma is in error.
 Rodriguez treats adverbs in two sections of the Arte; under the parts of speech (73v-77), and under the syntax (113-125). As has been observed in the introduction, there is little consistancy of classification between Rodriguez and Collado in this area of grammatical description.
 The interrogatives are derived from the Arte (110v) and are presented in substantially the same order. The adverbial particles which begin with uie are taken from (140-148v) and classified by Rodriguez as posposiao.
 The errata has; page 50, line 10, doco read coco. This would require the doco zo above to read coco zo. It seems that the errata should have read; page 50, line 16, which would have corrected this error. The punctuation is not corrected by the errata.
 Rodriguez has the complete version; Fitocuchi futacuchi cǔ cotoua cǔta vchideua nai.
 Rodriguez uses vonna for vonago.
 The material for this section is derived from the Arte (74v and 76v).
 Rodriguez has Ayamari nai vyeua, ...
 For the temporal interrogatives cf. Arte (89v-90v) and for the remaining forms 107-107v.
 Cf. the Dictionarium under cras.
 The Vocabulario has snuru and sannuru as the ombin form of the attributive perfective sarinuru.
 Cf. Arte (74v).
 The Spanish manuscript has iya iya.
 Cf. Arte (74v).
 Cf. Arte (75, 94v, and 123v-124v).
 Cf. Arte (94v) Quixoua ano fito fodono gacuxǒdeua nai.
 Cf. Arte (95 and 141).
 Cf. Arte (75).
 The Dictionarium has a selection of a dozen intensifying adverbs listed under valde.
 Cf. Arte (74v, 75, and 76).
 Cf. Arte (74, 75, and 75v).
 The Dictionarium also has the spelling moxi which suggests that Collado perceived a different vowel quantity than Rodriguez who has mǒxi, as does the Vocabulario.
 The Latin particle is nonne, which expects an affirmative answer.
 Rodriguez, and consequently Doi (Nihon daibunten, p. 449), have xidai for xisai. The original source is the Esopo no Fabulas where on p. 493 the form is xisai.
 While the material for this section has been drawn from various portions of the Arte, Rodriguez handles the bulk of the matters dealt with here on 106v-108v and 140-148v.
 The text is not clear at this point. It reads: Tame, significat ni vel erga: v.g. ... where one would expect: Tame vel ni significat erga: v.g. ... Ōtsuka translates this passage as if it were the later, as do I.
 Collado has recast into the colloquial a quote from the Shikimoku. Rodriguez records: Mata daiquanni itatteua ichininnomi sadamubequi nari.
 The text reads: itatte v.g. totte.... where the v.g. is clearly a misprint of vel.
 Cf. Arte (130-137).
 This item is the only one in this paragraph which Rodriguez does not list as a casane cotoba on 134v of the Arte. Collado is apparently interpreting this construction as a repetition of two adverbs, as for example coco caxico. If so, the form should be spelled vomoxir, vocaxi (if we follow his rule for the formation of adverbs from ij ending adjectives). However, the form which he seems to be recording is more likely the compound adverb which is listed in the Vocabulario as vomoxirovocaxǔ and glossed as contemporizando de boa maneira 'temporizing in a carefree manner.' The spelling that we suggest is derived from the attested lexical item without the application of Collado's formational rules.
 Cf. Arte (125-130v).
 This interjection, together with hat below, are the only uses of initial h found in the description. Rodriguez transcribes the latter item as at or vat (Arte, 127) which suggests a close relationship between the labial and glottal aspirates.
 Rodriguez has Benquei satemo yasaxij yatubaraya.
 Rodriguez has: ... nituaye bǒno saxivorosu. The entire passage would be, 'Benkei, seeing this, thought, "Oh, this isn't very important," and dropped the stick into the garden.' which Rodriguez explains to mean being sorry for not paying sufficient attention to a matter.
 The material for this section is derived from various sections in Book II of the Arte.
 Matthew, 6:24.
 Rodriguez has the spelling touazumba. In transcribing the form Collado failed to follow the rule he established in his treatment of conditional constructions.
 The model for this sentence appears to be Arte (62): Ichidan medzuraxij yenoco, que nagǒ, uquino gotoqu xirǒ [sic], me cur, cauo icanimo airaxijuo cureta. If this is the source of Collado's example, he is clearly demonstrating his sensitivity to the nasalization of such items such as nagǒ. The Dictionarium under longus has nagai.
 Collado's transcription is unable accurately to express the proper phonological, or morphological, form of shin'i 'indignation.' He would have been well advised to follow Rodriguez' model and transcribe this item as xiny with the specification that consonant plus y indicates a morphological juncture.
 Rodriguez has the spelling Quiso, which agrees with the Amakusaban Heike (p. 239), the ultimate source of the sentence. Collado's spelling in the translation is quiuzo. The Spanish manuscript has Kiso.
 One might expect the more literal 'I do not believe that it will be finished,' but Collado has credo quod non finietur.
 This rule, which might more appropriately have been included with the phonology, is not followed in Collado's description, with the possible exception of p. 48 where the same construction is apparently used.
 Collado here demonstrates the absorbitive capacity of Latin as he creates an accusative singular adjective from the past attributive of the verb kobu.
 The use of abiru, where one would expect aburu, may be a simple typographical error or evidence that Collado accepted the shift from ni-dan to ichi-dan katsuyō as unworthy of notice. Rodriguez (Arte, 101v) has midzuuo aburu.
 This list is derived from the Arte (101v-102v). From abi,uru on, the list is in the same order as that made by Rodriguez. Fanaruru, zzuru, nosquru, noru, vovaru, and mairu are Collado's contributions.
 Cf. Arte (101v).
 Cf. Arte (100).
 Cf. Arte (98).
 Cf. Arte (104).
 Cf. Arte (64 and 79).
 The material presented in this section is gleaned from the exhaustive treatment of the numerical system which makes up the last 20 leaves of Rodriguez' grammar.
 This compound does not follow the rule, since cu is not a iomi numeral. See also cu ninai below.
 Rodriguez has fitoi or fifitoi (Arte, 228v).
 While this form fits the general rule for combining counters and days, Rodriguez (Arte, 228v) has tǒca, which is a misprint for tca, cf. Doi, Daibunten, p. 818.
 Spelled with a tilde, sguat, as are all the other forms before guat.
 For the sǒ and sa allomorph of san cf. Arte (173v).
 Rodriguez gives the following equivalents in the monetary system on 217-217v of the Arte: ... ten Rin in one Fun, ten Fun in one Momme, one thousand Momme in one Quamme.
 The text is confused at this point. It runs: Ixxacu, unus palmus seu tertia quam Hispania vocant sanjacu. tres, ...
 The text has culus 'posterior,' but the errata changes the word to anus. The original seems closer to the Japanese.
 The examples here lag one behind the glosses.
 Here and elsewhere Collado combines homophonous enumerators which Rodriguez keeps distinct. Cf. Arte (220-223v) for an extensive list of enumerators.
 Cf. Arte (159-159v).
 This rule, apparently an invention of Collado's, has no precedent in Rodriguez or in linguistic derivation. The n in this construction is the contracted form of the classical mu, the source for what Collado calls the future.
 These forms might better have been presented as nari,i and qeri,i to indicate that the sentence-ending forms are nari and qeri.
* * * * *
Corrections made to printed original.
p. 14. 'BOOK II. The Rudamenta' corrected to 'Rudimenta'.
Ib. 'While the Arts Grammaticae presents ...' corrected to 'Ars Grammaticae'.
p. 16. 'booklet which teaches Canaduzcai' changed to much other uses as 'Canadzucai'.
p. 17. 'Tassuru, taxxita, taxxe ...' the last amended to match the paradigm 'taxxe'.
p. 19. 'Motone.' in future column, amended to fit the paradigm 'Motome'.
p. 20. Heading, 'Conjuctive' corrected to 'Conjunctive'.
p. 21. Table of irregular verbs, the left hand column has separate entries 'A' and 'Bi', these appear to mean a single entry 'Abi'.
p. 24. 'Those ending in u change to Tǒ' - 'to' omitted in text.
Ib. 'thus governs the genative' corrected to 'genitive'.
p. 116. 'There are ennumerable nouns' corrected to 'innumerable'.
p. 117. 'a visable thing' corrected to 'visible'.
p. 118. 'primative pronouns' corrected to 'primitive'.
p. 132. 'I did not decend.' corrected to 'descend'.
p. 136. 'vomi:vom', from the context and other references (and the Latin text) the root should be 'vomoi'.
p. 179. 'xi ban me forth' corrected to 'fourth'.
Footnote 27 a is applied to 'Nobi', this should be 'Vabi'.
Footnote 62. 'The term is also found in the introduction to the Vorabulario', corrected to 'Vocabulario'.
Footnote 106. 'verbo defectino' (from Portuguese text) corrected to 'defectiuo' as a more likely corruption than from 'defectivo'.
Footnote 109. 'selects that varient' corrected to 'variant'.
Footnote 131. 'the zzu varient' corrected to 'variant'.
Footnote 169. 'temperizing in a carefree manner' corrected to 'temporizing'.
Footnote 169. 'Spelled with a tilda' corrected to 'tilde'.