1. On the Oroch language and its history

The Oroch language is spoken in the Khabarovsk Krai, in far eastern Russia. According to the 2010 census, there are c. 600 individuals who declare themselves to be of Oroch ethnicity, and only eight individuals actively speak the language.

Although the Oroch people and their language have been known since 1781,Footnote ¹ proper linguistic data did not emerge until the very end of the nineteenth century. Before P.P. Schmidt published his materials in 1927 (they were collected in 1908),Footnote ² four sources containing Oroch data are generally acknowledged to have existed (see brief summaries in Gorcervskaja Reference Gorcevskaja1959: 22–3; Lar'kin Reference Lar'kin1964: 9–13; and Avrorin and Levedeva Reference Avrorin, Lebedeva and Skorik1968: 191).

Two new sources should be added: Arsen'ev (Reference Arsen'ev, Arsen'ev, Girfanova and Suxačov2007) and Tronson (Tronson A). The former was published only recently after the painstaking editorial work of A.X. Girfanova and N.L. Suxačov. Arsen'ev's field notes contain mostly Udihe linguistic and ethnological data, but some Oroch words can also be found. The latter has come down to us hidden in the pages of the otherwise largely unremarkable Personal Narrative of a Voyage to Japan, Kamtschatka, Siberia, Tartary, and Various Parts of Coast of China; in H.M.S. Barracouta, the recollections of a certain J.M. Tronson, a medical officer on board HMS Barracouta, published in 1859 in London. However, they were made public in a speech read before the Royal Asiatic Society on 15 April 1857 and published in 1859. A summary of known Oroch sources in given in Table 1.

Table 1

Summary of Oroch sources

4 The best way to consult Arsen'ev's Oroch data (see details below) is by looking at the index of Tungusic forms (pp. 266–427) where Oroch words are properly labelled.

Although Tronson's data has been known since 1861, the information has somehow been systematically ignored and assumed to contain Oroch materials. Tronson did not help matters as he misidentified the word-list as consisting of Nivkh.

Regrettably, the misidentification of Oroch linguistic materials has already occurred in the short history of this language. Lopatin's materials include glossaries in three languages: Oroch, Nanay (=Goldi) and, according to Lopatin, Ulcha. The Oroch–English glossary (or long list) is found on pp. 71–82, and the Ulcha–English glossary (or short list), on pp. 105–9. The latter consists mainly of Oroch, the correct identification being attributable to J. Benzing (Reference Benzing1958 = Johanson and Schönig Reference Johanson and Schönig1988: 53–9). Benzing (Reference Benzing1958: 600–1) provides five phonetic diagnostic features as well as some morphological formations which clearly set Oroch apart from Ulcha.Footnote ³

It is just as unfortunate that Oroch linguistic data are mistaken for Oroqen (Chinese 鄂伦春 èlúnchūn), which is in some instances seen as a highly aberrant variant of East Ewenki, and in others, an independent language (in both cases there is no discussion regarding its position as a member of the Northern Tungusic branch). This error originates in the somewhat confusing Russian terminology, according to which both ethnic groups are historically referred to by the name “Оро(н)чен/Оро(н)чён”. For instance, Skurlatov includes a brief Russian–Oroqen glossary (Skurlatovъ Reference Skurlatovъ1899: 254–62 sub “Ороньчонскiя слова”), which Clark (Reference Clark1981–82: 76 [94] = Clark and Walravens Reference Clark, Walravens and Walravens2006: 149 [120]) described as “Russian–Oroch”, though it clearly contains Oroqen data. “Orotchone”, as in the title of Leontovič and Protodiakonov's Oroch dictionaries, is the name used by Sabir (Reference Sabir1861: 78–83) to refer to the Oroqen.

2. The author: John M. Tronson

There is little information available concerning John Mortlock Tronson (1829/1830–1863). He passed RCS exams as naval surgeon, eventually becoming the acting assistant surgeon on HMS Barracouta (appointed 22 December 1852, see The Navy List 1853: 311) during its deployment to the Far East from 1854 to 1856. He is frequently mentioned with regard to military issues in the British press. According to his obituary, Tronson died in Sydney, Australia, at the age of 33. In a curious note, we are told (The Medical Times and Gazette 1863: 523, 525) that Tronson was thought to have passed away on HMS Orpheus, but in reality, as the note makes clear, he died at his residence.

As for his experiences on board HMS Barracouta, Tronson writes in the preface of his published travel recollections (London, 1859), that he visited “[…] various parts of the coast of China, the Japanese Islands, Kamtchatka, the coasts of Siberia and Tartary, and the island of Seghalien” in the years 1854, 1855 and 1856. By “the coasts of Siberia and Tartary” he is referring to the coastal regions of the Lower Amur, where he had the opportunity to meet local inhabitants.

There are linguistic remarks of the most diverse nature scattered throughout the book. Leaving aside the pseudo-Nivkh word-list, the rest of his remarks are not especially original or revealing.Footnote ⁵ Tronson collected specimens of Chinese, Japanese, Mongol and Manchu writing, which he later showed to various groups of natives in order to evaluate their proficiency in different languages. We are told, for example, that the Nivkhs did not recognize Chinese characters (p. 324), unlike the “Tartars”, who seemed familiar with them (p. 308); he shows Manchu writing to Koreans too, but they did not understand it. In spite of this, the Koreans requested copies of the Testaments in Japanese, which Tronson distributed among them under the petition of the Bishop of Victoria (p. 390).

There are also extensive annotations regarding various ethnographic aspects of the Nivkhs. Tronson read some of these observations before the Royal Asiatic Society in 1857 and they were published separately in 1859 together with what seems to be the same pseudo-Nivkh word-list.Footnote ⁶

Be that as it may, Tronson's personal interest in ethnographic and linguistic matters, and the opportunity he had to deliver some of his observations to an educated audience, makes his failure to distinguish Nivkh from Oroch even more surprising.

3. The word-list

There are two versions of the pseudo-Nivkh word-list by J.M. Tronson, both published in 1859. The first appeared in his monograph (Tronson A), the second was appended to a brief scholarly paper on the Nivkh read before the Royal Asiatic Society in 1857 (Tronson B). Though the lists are virtually the same, they show some interesting discrepancies which are relevant to the linguistic analysis below.

For the correct identification of Tronson's materials, we are indebted to the German–English geographer Ernst G. Ravenstein (1834–1913), who, perhaps out of politeness, decided to pass over in silence Tronson's blatant mistake regarding the linguistic identification of his own materials. Ravenstein's identification of Tronson's data as Oroch was possible thanks to the many reference works he had at hand (everything that had been published at that time). The Oroch were already well known by that name. Ravenstein may have been informed viva voce, or perhaps he inferred that Tronson was in fact referring to the Oroch via the books and articles of La Pérouse, Schrenk or, more likely, Lieutenant M. Veniukof, who in 1858 explored the surroundings of the Usuri river. Veniukof, whose observations are extensively quoted by Ravenstein, had the opportunity to meet the Oroch and described their homeland in great detail (Ravenstein Reference Ravenstein1861: 234–59).

Excellent summaries of scholarly work on “Paleoasiatic” (that is, Kamchukotic, Yukaghir, Nivkh and Eskaleut) and Tungusic languages such as Vdovin (Reference Vdovin1954) or Gorcevskaja (Reference Gorcevskaja1959), respectively, make no mention of Tronson: Vdovin perhaps knew that the language in Tronson's word-list is not Nivkh, whereas Gorcevskaja was perhaps unaware that it actually conceals Oroch linguistic data.

To the best of my knowledge, the first author after Ravenstein to notice that Tronson's word-list does not contain Nikvh data is R. Jakobson (Reference Jakobson1971[1957]: 74), who writes only that it is “[…] a list of Tungus words”, without going into detail regarding the exact linguistic affiliation of the words.

As for Tronson's surprising misidentification, the first clue that he is not on the right track may be his mistaken view of the difficulty of Nivkh: “Their language is simple, and easily acquired” (Tronson A 323–4) and “The language of the Ghilack is gutteral [sic], simple, and easily acquired” (Tronson B 117). As is well known, structurally speaking, Nivkh is the most complex language spoken in the Amur and Sakhalin regions. Even Šternberg, an accomplished linguist (at least a better one than Tronson), claimed that he was unsuccessful at mastering it after diligent study (Šternbergъ Reference Šternbergъ1908: viii–ix). Šternberg also notes that the Nivkhs did not speak their own language when approaching their neighbours, as no one seemed to have a good grasp of it, including the Russians who frequented Nivkh areas – instead they used Oroch, Ulcha, Nanay, Ainu or Russian.

Furthermore, in spite of his ethnographic interests, Tronson also failed to identify Nivkh traditions correctly. The most notable example concerns burial customs, about which Tronson writes the following:

This tradition, the so-called boat burial (also tree or aerial burial) is generally found among Tungus, not only in Sakhalin, but also in the Amur region and beyond. It has also been observed among the Yukaghir or the Yakut (see, for example, Lopatin Reference Lopatin1960: 117–9; on Oroch customs more specifically: “They employed the common Tungus forms of burial: on platforms set on piles, earth burial, and burial in a coffin in a burial hut”, Levin and Potapov Reference Levin and Potapov1964[1956]: 758. Nivkhs prefer cremation and, more rarely, inhumation (Lopatin Reference Lopatin1960: 120; on the Nivkh more specifically: “The burial rites of the Nivkhi differed sharply from those of the Tunguso-Manchurians of the Amur regions. The Nivkhi characteristically burned their dead, but they also buried them in the earth”, Levin and Potapov Reference Levin and Potapov1964 [1956]: 779–80. Even Ravenstein contains a passage on this very same fact: “The body is first burnt on a funeral-pyre, and a small wooden house is erected over the carefully-gathered ashes” (p. 392). Note that inhumation may be secondary to both the Oroch and Nivkh, due to Christian influences.Footnote ⁷

In spite of being surrounded by Tungusic groups, e.g. Oroch, Ulcha, Nanay, Udihe and even Orok in Sakhalin, the only time Tronson mentions them (“Tongoose”) is on the very same page as the Nivkh word-list. Ironically, in Tronson B (118) he writes: “I have not been fortunate enough to meet with the Tongoose, though I visited some of their huts in the summer of 1855” and “I learned from a Russian officer, that the Tongoose, unlike the Ghilack, are cruel and treacherous”.

4. Linguistic commentary on Tronson's word-lists

Linguistic data in Tronson A and Tronson B are organized and presented in two-column word-lists which are found on p. 325 and p. 119, respectively. Both word-lists contain 45 words: 35 regular lexical items and 1–10 numbers. Tronson writes: “I append a few words in the Ghilack tongue,Footnote ⁸ which I obtained in learning to converse with the people. They are very anxious to give the name of everything when asked, but could make very little progress in the acquisition of the English tongue” (Tronson B 117), therefore we are dealing with genuine, original linguistic data, if Tronson can be trusted.

Tronson A contains no special symbols, apart from the macron and the acute accent. Tronson B, however, contains umlaut in at least two instances (see below [1] and [26]) and grave accents instead of acute accents. Given the lack of special symbols in Tronson A, one can surmise that Tronson sought to follow closely English spelling conventions, so that casual readers could easily make sense of his annotations.

Generally speaking, although impressionistic and not always very consistent, nineteenth-century Oroch sources are very reliable and the only effort required on the specialist's behalf is to analyse carefully the transcription used by their authors, which is amateurish and always very close to the spelling conventions applied to their native language.Footnote ⁹ This is a common trait in Tungusic pre-scientific sources and, as the analysis below will show, both remarks (on reliability and transcriptional problems) hold true also for Tronson's word-lists. The overall impression is that Tronson B is more reliable than Tronson A. There are, however, five major discrepancies, see individual comments below for items [4, 7, 18, 26, 27].

Table 2 shows Tronson's 45 words set against the same words as found in other Oroch sources. Tronson A and B are reproduced when items are written differently. For the sake of comparison, modern forms of Oroch and Nivkh equivalents are also included.Footnote ¹⁰

Table 2

Comparative vocabulary list of Tronson's word-lists with other Oroch sources and Nivkh

Some of the items in Table 2 demand individual comments:

13 In the collocation waca inaki ‘a bitch, a female dog’ (60a).

14 As for the orthography of this word, there is a third version in the body of the text in Tronson B: “In Destination Bay, the god was in the resemblance of a man, and in order to witness the effect of its removal on the minds of the people, I pretended to remove it; they were indignant and horror stricken at my temerity, pointing to the heavens. They clasped their hands and bowed their heads saying, “Nang-Nha! Nang-Nha!” They gave me to understand, that if I removied it, sorrow and woe would be their portion in this world and in the next” (p. 114).

15 Nivkh distinguishes numerals for people / for animals.

16 [43–44] in Margaritov are highly deviant forms. <g> in xugó ‘9’ may reflect the fortition of a (secondary) epenthetic element -w- between labial vowels, rather than a very archaizing pronounciation, cf. PT *hägün (though PT *-g- > Oroch Ø, perhaps remodelled on Amurian Ewenki = Literary Ewenki yägin). As for ‘8’, -t- may perhaps be better seen as the remnant of the cluster -kt-, which is documented only twice, namely in the collocation japtu datki = †japtü daakti ‘eight rows, layers’ (TMS 252a, cf. Ulcha jakpündaktï id.) and <dzjaktu> ‘eight’ (Schmidt 32a, collected by Brailovski near the Koppi river, see Map 2), the origin of which remains unclear.

1. [2] Šternberg meuci (445a). This is a well-known Kulturwort which spread across Lower Amur and Sakhalin from Manchu mioocan/m'oucan/, which in turn comes from Chinese 鸟枪 niǎoqiāng ‘shot gun, air gun’ (Schmidt Reference Schmidt1933: 370a, TMS 1.562–3, Doerfer [7057]),Footnote ¹¹ cf. Udihe niinka(m) m'aosá ‘Chinese handgun’ (Arsen'ev Reference Arsen'ev, Arsen'ev, Girfanova and Suxačov2007: 207–8, 434). As for the Nivkh form (← Tungusic, see Krejnovič Reference Krejnovič1955: 160), it is documented in Furet (Reference Furet1860: 109 <Miotché>, see also Grube Reference Grube1892: 127 sub Flinte).

2. [4] Tronson B shows that <M> in Tronson A is a clerical error. In this case, as in others, it is impossible to assert whether the typographical error was committed by the publisher, or whether the publisher was dutifully transcribing an error committed by Tronson himself.

3. [6] Šternberg (Reference Šternberg and Šternberg1933: 441b) and Arsen'ev (Reference Arsen'ev, Arsen'ev, Girfanova and Suxačov2007: 431) collected and described the names of various subtypes of untá which vary in function and the skin they are made of. See illustration in Lar'kin (Reference Lar'kin1964: 64)

4. [7] Tronson B shows the singular, regular form, whereas Tronson A may be a transcription of a very archaic formation containing the plural(izer, collective) ending +l ~ +r (see Benzing 1956: 75), cf. Solon bäldiir, lit. Ewen böödŭl, Arman budäl, etc. (TMS 1.118–9). Although there may be some relics in Southern Tungusic,Footnote ¹² this is a proper Northern Tungusic formation, where the regular plural marker is +l (unlike in Oroch, where we only find +sa). Cincius (Reference Cincius1949: 121–2) was the first scholar to notice some lexical parallels existing between (Okhotsk) Ewen and Oroch (sometimes Udihe too). Those, however, can be accounted for by assuming that they are the result of sharing a common Amurian culture. There is no special reason why leg(s) would have to belong to that category. Notwithstanding this, cf. [33].

5. [9] This is most likely (a dialectal variation of?) Chinese 鲶鱼 niányú ‘catfish’ (already in Maak Reference Maakъ1861: 194–5). Catfish (Nivkh mupïk) are extremely common in the Amur region, and this may be the reason why it appears to be the regular word for fish (see below [10]).

6. [10] sugjasa is the regular word for fish.

7. [11] The regular word for deer is oro. Tronson's †talŋi bears an obvious resemblance to Sakhalin Nivkh tlaŋi ~ Amurian Nivkh c^holŋi ‘deer’ (Grube Reference Grube1892: 79, 124; Savel'eva and Taksami Reference Savel'eva and Taksami1970: 451b). Nivkh tlaŋi seems to have a tendency to travel across the sea, as Austerlitz (Reference Austerlitz, Heissig, Krueger, Oinas and Schütz1976), after proposing a convincing internal etymology, argues that the same word was loaned to Sakhalin Ainu, where it took the form tunakai ‘reindeer’, and from there, via Hokkaido, become a part of the Japanese vocabulary as トナカイ tonakai.Footnote ¹⁷ This word is recorded in two Santan (Ulcha) word-lists dating from 1882 and 1886, though linguistic data probably goes back to the Tokugawa period (Ikegami Reference Jirō1967: 37, 83). As for its presence in Oroch, we have to assume that the word spread westwards when the Nivkh visited the Lower Amur for trade. Such contacts are well documented (see Figure 1).

   Figure 1.

   Linguistic interaction in the Lower Amur and Sakhalin (based on Wurm et al. Reference Wurm and Peter1996, vol. 1, map 108. Russian is not included.) Dotted arrows indicate “passive bilingualism”)

   

8. [13] ñaŋña means ‘heaven, (visible, clear) sky’. The regular word for god, spirit is änduri (← Manchu enduri ‘spirit, god, deity’) or büa ‘place, world; sky, weather; (superior) deity, spirit’ (Avrorin and Lebedeva Reference Avrorin, Lebedeva, Avrorin and Lebedeva1978: 168b; Šternberg Reference Šternbergъ1908: 447a sub Religija). *ñ- is original (< PT *ñäŋñä, see TMS 1/653–4; Doerfer Reference Doerfer and Doerfer2004: [8049]), therefore the fact that it has not been recorded by Tronson and Leontovič may be due to poor language skills or to an idiosyncrasy in the informant's speech (sporadic dissimilation?), cf. [41].

9. [18] The English translation ‘canoe’ (cf. Schmidt Reference Schmidt and Schmidt1928: [61a] ‘a large river-boat’) in Tronson B is the correct one, ‘cause’ in Tronson A is obviously a typo. This is an obvious derivative of uli ‘river’, cf. [28]. Lar'kin (Reference Lar'kin1964: 38–9), Šternberg ulmagda (444a) and Arsen'ev ulimagda (119).

10. [20] The same as Ulcha xacoa ~ xacüa, Orok xacuga (=? Ikegami Reference Jirō1997: 228 s.v. xacuwa ‘water container’), Nanay xacoxã, and Kili xacãã ~ xaoxoã. All of them mean ‘(hanging) cauldron’ and go back to Manchu hacuhan ‘a small cooking pot’ (see TMS 2.464–5; the loss of *-k-, preserved in Manchu as -h-, is self-evident in Oroch, Kili xacãã and Ulcha). Šternberg's xaoca ‘(kitchen) utensils that may be used as plates’ (443a) is certainly the same word. The vocalism in the first syllable in Kili and Šternberg's form has perhaps been remodelled on a Chinese word, e.g. 陶器 táoqì ‘crockery’ or 好吃 hǎochī ‘nice, delicious’.

11. [23] The h-less form in Tronson B may have been a slip of the memory which stood corrected a posteriori in Tronson A, or simply a clerical error.

12. [24] Tronson seems to have recorded the Ulcha word gïa ‘jaw(-bone)’, which along with Nanay geã <гиан>, is characteristic for lacking the diminutive ending *+kaan, as in Lit. Ewenki geekaan, Solon giixaa, Kili gïkã or the proper Oroch form (< CT *gya(a)+kaan, see TMS 1.145b, Doerfer Reference Doerfer and Doerfer2004: [3949]). Sub подбородок ‘chin’ and борода ‘beard’, Leontovič has guggaxta (48, 105), which is the same word as Protodiakonov kukäxta ‘moustache’ ~ kukäxta ‘beard’ (46) and güggakta ‘beard, moustache’ (Avrorin and Lebedeva Reference Avrorin, Lebedeva, Avrorin and Lebedeva1978: 174b), cf. Udihe budzjakta ‘the chin, the beard’ in Schmidt (27b). Margaritov has guaxtá ‘beard’ ~ guaktá ‘moustache’ (47, 52 ← Udihe guakta ‘beard, moustache’).

13. [26] Tronson B has the correct form with <l> instead of the one with <B> in Tronson A, most likely a clerical error.

14. [30] Apart from ‘cape, promontory (Russian мыс)’, Avrorin and Lebedeva Reference Avrorin, Lebedeva, Avrorin and Lebedeva1978: (179b) provide the additional meaning of ‘the end of any long object, path or territory’. Cf. Avrorin and Lebedeva xuwä ‘taiga, forest’ (248a), Nivkh pal ‘taiga, forest’. Margaritov dïgí ‘forest’ (49) and Schmidt dihi ‘wood, forest’ (30b) are perhaps the same as the adverbials diixi (Xadi) ~ diisi (Tumñin) ‘uphill from the coast; up, upward(s); in the taiga; to the west’ (Avrorin and Lebedeva Reference Avrorin, Lebedeva, Avrorin and Lebedeva1978: 178b), from which it would be easy to gain the impression that the word actually means ‘forest’.

15. [28] If this actually reflects uli, then <dh> (= d) remains unaccounted for.

16. [31] The retention of *-g in Auslaut is a consistent feature of Northern Tungusic, cf. Lit. Ewenki & Solon & Negidal beega, Lit. Ewen bëëg, Arman beek. In other words, the loss of *-g(-) characterizes Southern Tungusic, e.g. Ulcha bee, Orok bee, Udihe beë, Lit. Nanay bea <биа>, Manchu biya /b^ya/ (< PT *bya(a)ga, see TMS 1/78–9, Doerfer Reference Doerfer and Doerfer2004: [1231]).

17. [32] Tronson himself did not know Russian. Therefore, he could have misheard the Russian word внучка ‘granddaughter’ or even perhaps one of the affective-diminutive suffixes -очка ~ -ушка, as in девочка ‘(little) girl’ ~ девушка ‘(unmarried) girl’, and assumed that this is actually the word for child.

18. [33] If Tronson's transcriptions render something like †bäsäg(d), there are potential matches in (Okhotsk) Ewen bäysäg ‘group of people; family’ and bäysäc ‘snob, big-head’ (both are derivates of bäy ‘person’ < PT *bäyä, see TMS 1.122–3, Doerfer Reference Doerfer and Doerfer2004: [1415]),Footnote ¹⁸ which seems to lack exact parallels in other Tungusic languages. Moreover, PT *bäyä continues in Southern Tungusic with the meaning ‘body, -self’ and is barely present in derivative formations. <dh> (= d elsewhere) remains unaccounted for.

19. [34] Šternberg Reference Šternberg and Šternberg1933 aduli (446a). Tronson B does not contain the awkward symbol <!>, which may be just a clerical error that occurred during the typesetting of the text.

20. [36] Most likely ommoko ~ ämäkä, omokko ~ ämäkkä ‘only one’ (TMS 2.271a, Avrorin and Lebedeva 215b, 257b), a form which contains the regular cardinal numeral omo ~ omu ‘one’ and the so-called restrictive suffix +(k)ka (Avrorin and Boldyrev Reference Avrorin and Boldyrev2001: 385–6). See also Lopatin omokú ‘alone, single, solitary’ (Reference Lopatin1957: 78b), cf. Ulcha omokä(n-) ~ umukä(n-) ‘solitary, lone(ly)’, etc.

21. [41] *ñ- is original (< PT *ñöŋön, see TMS 1.647–8, Doerfer Reference Doerfer and Doerfer2004: [8115]), therefore the fact that it has not been recorded by Tronson and Protodiakonov may be due to an idiosyncrasy in the informant's speech (sequential analogy to nada ‘7’?), cf. [13].

If the linguistic analysis provided above is correct, Tronson's word-list contains a series of lexical items showing very clearly the effects of both language contact and historical evolution. Given the ethnic configuration of the Lower Amur region during the last centuries (see, for example, Janhunen Reference Janhunen1996: 113–7; Khasanova Reference Khasanova, Gilbers, Nerbonne and Schaeken2000; Zgusta Reference Zgusta2015: 104–4), it is rather natural to find one word from Ulcha ([24]), one from Nivkh ([11]), one from Russian ([32]), one from Manchu ([20]), and two Chinese words, one straight from Chinese ([9]) and another via Manchu ([2]). The Nivkh loan is, however, somewhat special, as it is documented in a Tungusic language for the first time.

Furthermore, there are at least three items which undeniably possess Northern Tungusic pedigree ([7, 31, 33]). These have never before been observed in Oroch. If anything, they only add to the evidence supporting the current view that sees Oroch as a (peripheral) Northern Tungusic language which has undergone some changes under the influence of Southern Tungusic, especially Nanay and Ulcha (Georg Reference Georg and Näher2004). With this evidence at hand, it is possible to argue that Tronson is the only testimony confirming the presence of certain Northern Tungusic features in Oroch which were replaced under the influence of Southern Tungusic between the 1860s and the 1890s.