The Marakovic Collection
Hugh Moss and Stuart Sargent


Part I: The Imperial Phase
The Kangxi Era, 1661-1723

Tobacco and snuff, introduced from the Americas via Europe, began to take hold in China in the first half of the seventeenth century. Although some Westerners on the fringes of the empire may have been snuff takers, along with a few of the Chinese who came into contact with them, it seems from early references that the overwhelming preference was initially for smoked tobacco. This legacy continued throughout the dynasty among the lower classes.

It was the Kangxi emperor who promoted snuffing as a replacement for smoking. He accepted gifts of imported snuff and bestowed both snuff and snuff bottles on those whom he wished to honour. In 1684, during the first of the southern tours he undertook to assuage the animosity of a population that had suffered greatly during the Qing conquest, he accepted a gift of snuff from missionaries in Nanjing (while rejecting a series of other proffered treasures). He was already thirty years old then, and appeared to have prior knowledge of snuff. In one of the principles of moral conduct that he set down (later collected and published by one of his many sons, the Yongzheng emperor), he said that although he was ‘good at smoking’ when he was a youngster in his wet nurse’s quarters, he was nevertheless giving it up forever because he was prohibiting the practice as a fire hazard, and he could not ask others to do something he himself was not willing to do. According to a nineteenth-century account, the Kangxi emperor also considered tobacco to be a health hazard; a different work, completed in the Jiaqing era (1796 – 1821), reports that he considered it a crop that should not be permitted on lands better used for food production. Of course, the Kangxi emperor was neither the first nor the last ruler whose influence over his subjects proved no match for nicotine.

The snuff bottle would have evolved naturally and quickly. Although some Chinese at southern seaports may have seen snuff boxes from Europe, and snuff boxes were presented to the emperor by papal envoys in 1705, the Chinese traditionally stored anything that needed to be kept fresh in bottles of some type that could be sealed or stoppered. In addition, snuff from abroad initially seems to have been delivered to China in stoppered glass jars. It is possible that the Chinese occasionally took smaller jars made for other uses and converted them to contain snuff. Small bottles of sizes and forms similar to snuff bottles already existed: they have been found in tombs dating from long before snuff was introduced to China.

As soon as snuff took hold at court, however, and in consideration of the fact that good-quality snuff was exclusively an exotic and expensive import, common, all-purpose containers would have been a clear mismatch. Far better to convert earlier carvings in jade (amulets, paper-weights, etc.), which were already a higher art form among the influential minority, or to make bottles specifically to function simultaneously as containers for snuff and as art. The court could call on its palace workshops, which were expanding rapidly in range and output during the very period when snuff was taking hold. Private workshops around the capital and further afield could also be enlisted to supply the newly addicted northern elite. Before long, any impulse to use earlier containers would have become an anomalous exception.

The snuff bottle as such was probably a development of the first twenty years of the Kangxi era. Glass may not have been the earliest material used to make snuff bottles for court use; in fact, given the known date of the establishment of the imperial glassworks late in 1696, it is unlikely. But glass is one of the few materials where we have any evidence at all to which we can cling in identifying Kangxi snuff bottles.

A famous and oft-quoted passage by Wang Shizhen 王士禎 (1634 – 1711) from a collection of jottings he made after returning to the capital at the beginning of 1703 specifies only glass bottles (ping 瓶) produced under the Imperial Household Department (i.e., in the imperial glassworks) as the storage vessels used for snuff, but he must be talking about what we call snuff bottles (biyan hu 鼻煙壺). Wang’s list of colours, ‘red, purple, yellow, white, black, and green’ is probably a list of general rubrics under which more specific hues could be given, for the archival records of the following era contain twenty-nine colour terms (not counting combinations in overlay and other techniques). He singles out two colours as particularly attractive: ‘clear like crystal; pink like pearl’, bai ru shuijing, hong ru huoqi 白如水晶,紅如火齊. The second colour requires some discussion. Hong can mean ‘pink’ or ‘red’; huoqi is a rare term variously explained as referring to pearls or rose-coloured mica (lepidolite or lithium mica) imported from (or more likely, through) India. Huoqi may even be a foreign word whose pronunciation could be recorded in Chinese with the characters huo and qi; the fact that this combination coincidentally denoted ‘the timing and strength of fire’ in ancient texts is irrelevant in this context, but because it includes the character for ‘fire’, it nicely parallels ‘crystal’, which contains the character for ‘water’; this is probably why Wang Shizhen seized on huoqi to describe what must have been a lustrous pink, probably derived from colloidal gold and part of the new palette of colours of glass introduced by the missionaries in 1696. He may have had in mind anything from deep ruby red to a series of paler colours, all derived from the gold.

Wang Shizhen’s list of colours is also interesting for its omission of blue; transparent sapphire blue is found among the broader range of wares thought to be Kangxi, considered to be of European manufacturing style, and observed to suffer from glass disease and crizzling. Conversely, Wang mentions green ( 綠), but transparent green glass attributable to the Kangxi era is conspicuous by its rarity. It may be that sapphire blue belonged under the general rubric of ‘purple’ (zi 紫) in Wang’s thinking, and that his ‘green’ was some kind of turquoise. We must accept that today we cannot be sure that Wang Shizhen saw colours the same way his contemporaries did, let alone the way we see them today. In any case, what is demonstrated beyond a doubt by this valuable source is that by 1703 the imperial glassworks was making a fairly wide range of colours, and that these were used in snuff-bottle production.

The Marakovic collection is not only strong in glass bottles, it contains the only known monochrome glass snuff bottle with a credible Kangxi reign mark (fig. I.1). This is one of but a tiny number of known Kangxi-marked glass wares of any form. Many other bottles may be candidates for a late seventeenth- or very early eighteenth-century date, but very few of them can be so assuredly identified as a Kangxi product.

The Marakovic bottle can be confidently dated to a limited period between the founding of the imperial glassworks under the direction of the Bavarian Jesuit Kilian Stumpf late in 1696 and the end of the Kangxi era in early 1723. (The Kangxi emperor died on 10 December 1722, but the new era on the calendar, following standard protocol, did not commence until the first day of the new lunar year, which was 4 February 1723 on the Western calendar.) The bottle is likely to have been made closer to 1696 than to 1723, but we cannot be certain of this. It is of sapphire-blue glass, one of the early staples, with relatively crude engraving suggesting experimental work and, possibly, Jesuit involvement. (The engraving and faceting of glass would have been among the glassmaking skills of the missionaries). The simple two-character mark suggests a date before the last decade of the era, by which time the few examples of Kangxi reign-marked wares from the palace had evolved through seal-script (zhuanshu 篆書) marks to four-character marks in regular script (kaishu 楷書). A two-character mark in regular script is, therefore, likely to be among the earlier experimental stages in this evolution, or perhaps mid-way, but not from the final phase, which probably equates to the last decade of the era.

The Marakovic bottle appears to have come from a tomb at some point; burial seems to be responsible for most of the surface degradation. Transparent glass from the early phase of the palace glassworks was usually extensively crizzled and sometimes suffered from glass disease, both conditions being due to improper chemical balance. In this case, the blue glass is not entirely transparent. With the stopper removed, one can see that the glass is in two layers, both of which seem to involve some opacifying agent. It is well known that whatever opacifying agent was used at the time resolved the problems of glass disease and extensive crizzling. Thus, white glass enamelled in the Kangxi reign is unaffected by either problem, and translucent or opaque glass from the Qing dynasty is very rarely compromised; when it is, it is usually as the result of long burial rather than any inherent problem of glass chemistry. That there is surface degradation on this bottle despite the use of an opacifying agent must be due to something in the soil, not to the composition of the glass itself.

The inscription in figure I.1 is the first four lines from no. 161 in the Book of Poetry (an anthology that assumed its present form around 600 BCE): 呦呦鹿鳴,食野之苹。我有嘉賓,鼓瑟吹笙。The translation of James Legge, given below, accurately reflects the fact that this is a feasting song and a celebration of fellowship.

With pleased sounds the deer call to one another,
Eating the celery of the fields.
I have here admirable guests;
The lutes are struck, and the organ is blown [for them].

(In an allusion to the first line, the guest house and reception hall for foreign guests that was built in the 1880s in Tokyo was called the Rokumeikan 鹿鳴館, the Hall of the Calling Deer. It is doubtful that any of the Western guests who attended balls and parties there understood the name.)

The Marakovic collection is strong in faceted bottles. Faceting was popular as a court staple throughout the dynasty but evolved under the years of Stumpf’s directorship of the glassworks, between 1696 and 1720. A Bavarian technique, the faceting of glass appears to have been introduced by Stumpf and his associates working at the court. The technology involved a large spinning metal disc to which abrasives were added; the vessel to be faceted was held against it to flatten the chosen area. Glass bottles in general were blown into moulds, so the first stage in some cases may have been to blow the glass into a mould that had the rough shape of the finished vessel, including flattened panels. We can be certain that this was the case with some examples where surface fragments of differently coloured glass occur evenly across all facets (see, for instance, fig. I.32). If the blown form had been bulbous, faceting the bottle would have removed these fragments from the main area of each flat panel. With a bottle blown into a mould with facets, the surface fragments are largely preserved, the faceting wheel being needed only to bring greater integrity to the facets and to the edges where they met—to sharpen it all up a bit. A further reason to think that this was the standard method is that blowing (with or without a mould) a basic form such as a flattened sphere or ovoid, leaving the glass thick enough so that it could then be faceted against the disc, and then grinding away a good deal of the glass to make the facets would have been an inefficient use of time, glass, and abrasives.

No known faceted bottles bear Kangxi reign marks, but this does not mean they were not a staple of Kangxi production. As we have noted, reign-marked wares were the exception to Kangxi imperial production, not the rule. There is credible evidence of the imperial production of faceted octagonal bottles that comes from early in the Yongzheng era, which followed the Kangxi era and spanned the years 1723 – 1736. On 28 January 1726 (to facilitate consulting primary sources, we shall also give the Chinese dates, where appropriate, starting with an abbreviation for the era and the year of the era, and then giving the month and day, so this date, the 28th day of the 12th month of the 3rd year of the Yongzheng era, is expressed as YZ3/12/28), a ‘grape-coloured, eight-angle snuff bottle’ was delivered to the imperial glassworks. With it came these instructions: ‘using this as a model, make several bottles in various colours’. The model was undoubtedly in the standard faceted octagonal shape we know to have been a staple of the palace workshops starting in the early eighteenth century. The ‘grape colour’ of the glass presumably refers to the colour purple. The bottle sent to the glassworks as a model in early 1726 may have been a year or two old, possibly old enough to date from the end of the Kangxi era. There are many mentions of earlier bottles being sent in for copying or repairing during the early Yongzheng period, even from the first year; these can only be bottles produced at court during the Kangxi reign.

Several other records from the Yongzheng palace workshops refer to glass snuff bottles ‘with corners’ being made in some quantities. The word we translate ‘corners,’ leng 楞, could be a synonym for ‘angle,’ as in ‘eight-angle snuff bottle’—in which case we are talking about our faceted bottles. A second possibility is that these are six- or eight-sided bottles that do not necessarily have flat facets and sharp corners. A third, less likely, possibility is that leng designates rounded lobes running vertically on a vessel’s body—gua leng 瓜楞, ‘melon lobes’, is the usual term for this. Lobed bottles are a standard palace shape and may have begun in the Kangxi period, but given the profusion of faceted octagonal ones, it is more likely that the ‘eight corners’ refer to them, not to lobed forms. Other glass bottles are said in the Yongzheng records to be ‘without corners’; these must be gently curving forms.

Whatever the precise shape meaning of ‘with leng’, the records tell us unequivocally that such bottles were made in ‘various colours’. Court production for the era is characterized by a fascination with colour, with close to thirty glass colours recorded for the Yongzheng glassworks, as noted above. We cannot know whether all of them were used for snuff bottles, but a wide variety of colours is represented in the Marakovic collection.

The unmarked bottles in the collection can be ascribed to the Yongzheng, Qianlong, or even later reigns, but some among them may have been produced during the Kangxi period. If there had been a smooth progression from a high level of crizzling and glass disease to a low level, we could use those factors as a way to date unmarked bottles, but the evidence is simply too shallow to posit such a progression. During the quarter century between the establishment of the glassworks and the end of the Kangxi reign, it is quite possible that they got the problem under control, but there could have been lapses. And, of course, translucent glass (allowing the transmission of light but not transparent) or the rarer opaque glass (where light is completely blocked), could have been produced without any crizzling at any early date, because they both included the protective opacifying agents (see our comments on the bottle in fig. I.1, above).

Figure I.2 shows one example of the yellow faceted bottles that are not of the well-controlled imperial-yellow colour widely produced in the post-Kangxi period. In light of the strange mix of colour and what seems to be a slight degradation of the exposed surfaces, which are paler than the rest, this multi-faceted bottle might be relatively early. Once various colours of imperial yellow were well understood, such minor problems would have disappeared.

Figure I.3 represents one of the other standard forms for these faceted bottles, that is, an octagonal profile whose two main sides have a raised convex panel that is either 1) of faceted, roughly triangular shapes meeting as a point at the centre, as here; 2) convex without facets; or 3) flat. This bottle is imperial yellow, quite possible for the Kangxi period. But since we know this crisply cut, formally symmetrical, long-necked, wide-mouthed shape was a standard imperial form that continued into the Qianlong and possibly even beyond, we have no way of distinguishing Kangxi from Yongzheng or Qianlong examples unless they are marked, which is extremely rare.

Figure I.4 would seem to show a likely early product, as the surface is somewhat degraded in places, but (as with the bottle in fig. I.1) this is due to burial and variations in how the bottle was cleaned and polished after it was excavated; no particular age can be deduced from the condition of the surface in such circumstances. The bottle in figure I.5, of transparent sapphire-blue glass, appears convincingly early, but it is perhaps unlikely to be from the Kangxi era, since it is not crizzled or diseased. Again, however, this kind of evidence may not tell us anything at all. Some Yongzheng glass of this range is not crizzled, so why should some Kangxi glass not be free of problems, too? The blue here is somewhat dulled or frosted, but this may be from the tomb rather than from any original chemical shortcoming.

All the same logic could be applied to the bottle in figure I.6, an example with somewhat frosted, colourless glass; it might be as early as the Kangxi but is certainly eighteenth century in any case. Another colourless example, less dull but with obvious surface signs of having been excavated, is illustrated in figure I.6a. Similar Marakovic bottles in turquoise blue, pink, ruby red, white, two different translucent green colours resembling jasper, and dark olive green (figs. I.7-I.8-I.9-I.10- I.11) are also possibly from the Kangxi era, but they are more likely to date from later in the eighteenth century.

Many other faceted forms may also be from the earlier phase, including the rare variation on the standard form with an additional register of faceting between the main-side panels and the narrow sides illustrated in figure I.11a. Again, the lack of crizzling on amber-brown glass suggests the likelihood of a date later than the Kangxi, while the wide mouth and crisp faceting suggest an eighteenth-century bottle.

The recent treasure trove of bottles excavated from private tombs during the expansion of Beijing over the past half-century has provided us with a mass of fresh data to sift. These bottles are dominated by recognizably imperial types, indicating that they were bestowed upon worthy individuals as signs of imperial favour (or as containers for gifts of precious snuff that signified the same favour). When moving on to the great bureaucracy of the afterlife, one would not want to tote an anonymous compressed spherical bottle in a colour of no particular significance imperially when one could take along instead a faceted octagonal (occasionally decagonal, but the impression would remain unchanged) bottle that would immediately be recognized, regardless of its colour, as an imperial product.

The same prestige principle applied, probably, to other well-known imperial forms, like that of the bottle in figure I.12, which is unquestionably early, although perhaps more likely on a simple ‘actuarial’ basis to date from the Yongzheng or early Qianlong era than from the Kangxi (the earlier the bottle, the more vicissitudes it must weather). Figure I.13 shows a typical early double gourd made at the court. It is of a type that was rarely found before excavations began in Beijing. Colourless and other transparent examples are extant without crizzling, suggesting that this crisply made, confident, wide-mouthed shape, which was presumably developed during the Kangxi era, continued to be an option for snuff bottles as glass chemistry improved. This turquoise example probably dates from a little after the Kangxi reign.

The bottle in figure I.14 almost certainly does not date from the Kangxi era, but black is one of the colours that Wang Shizhen mentions in 1703. Such a bottle as this is probably mid Qing, but with so plain a glass bottle, which has most likely been re-polished occasionally during its lifetime (the last time probably quite recently), clues to dating are as opaque as the material. The fact is that the glassworks produced an enormous number of bottles every year, certainly into the early nineteenth century and to some extent to the end of the dynasty, bottles that were almost invariably undecorated (though they might be monochrome, speckled, striped, swirled, sandwiched, or made to imitate various other materials, including realgar glass) and, being undecorated, practically impossible to date with any confidence. The bottle in figure I.15 may be considered a token example of this massive range of bottles. Whether the wide mouth is a reflection of the standard for enamels during the Kangxi reign or a mid-Qing fashion statement, we have no way of knowing, but it is one of many bottles from this group that may date from as early as the Kangxi reign – as some of them must!

Nino Marakovic has had a long-standing appreciation of cloisonné, so when he began to collect snuff bottles, this group provided a natural focus for him. He has formed one of the most significant groups of cloisonné bottles in the world.

Cloisonné wares were produced in one of the earlier imperial workshops; current scholarship has it predating the glassworks, which would not be surprising, since cloisonné was a traditional craft known to the Chinese since the early Ming dynasty. Although glass was produced in China centuries before that, a full glass technology including the capacity to blow relatively large pieces with metal blowpipes seems to have been introduced only late in 1696. The cloisonné workshop, on the other hand, was possibly in production as early as the 1680s. When we see references to enamelling in the 1690s, they must be to cloisonné; they can hardly be related to painted enamels on metal or glass, which were not introduced by missionaries until probably around 1700, give or take a few years.

Known Kangxi-marked cloisonné vessels from the palace workshops are notable for the absence of imported enamels. They seem to use the traditional palette exclusively (or at least predominantly). The wire-work is also distinctive in many cases, with carefully hand-wrought wires of varying thickness; later wares tend to be made with wires of the same thickness throughout. However, if the Kangxi workshops were in production for thirty or forty years, it is possible that there could have been an evolution in technique during those decades. Snuff bottles with more even wire-work and the inclusion of enamels from the imported palette, which became known as the famille rose palette in European publications, could still be from the Kangxi workshops, just from later in the era.

It might be misleading to imagine an abrupt transition: it is possible that the adoption of the new palette of enamels in cloisonné wares was gradual because this was a traditional craft with a well-established palette; there was no urgency to change what was already accepted, especially when the scarce imported enamels needed to be conserved for the new arts of painting with enamels on metal and glass and for decorating porcelain vessels sent from the kilns of Jingdezhen 景德鎮. We know from the records that these enamels were in short supply relative to demand during the late Kangxi and early Yongzheng reigns. This would have remained the case until 1728, when the European enamels were finally replicated at the palace workshops.

One problem in our efforts to identify early palace products is that, while we can reasonably judge whether a particular bottle or other vessel is early and experimental, we cannot necessarily be certain that it is early, experimental, and from the palace workshops, rather than from some other manufactory struggling to master the art. Although the art of cloisonné had been practiced since the early Ming dynasty, the palace workshops of the early Qing were new and relatively inexperienced, so they would have had to master the art again, just like any private workshop.

There are several apparently early and obviously experimental works in the collection, the first two with none of the famille rose enamels imported from Europe (figs. I.16 and I.17). This suggests that they were made before such enamels were available or at a time when they were too much in demand for enamelling on glass and for painted enamels on metal to be used on the nascent cloisonné snuff bottles. If they are from the imperial workshops, they may date from the period between 1680 and about 1715, although it is possible that other northern workshops without access to the new enamels began experimenting with snuff bottles at an early date. Both bottles have mouths whose width is a natural consequence of the construction of the bottles, and both exhibit the relatively crude use of traditional enamels with the sort of pitting one might expect of a start-up workshop. Neither is enamelled on the inside, although I.17 seems to have some patches of enamel. A lack of interior enamels would have been typical of early cloisonné snuff bottles. Bare-metal interiors were also a feature of early painted enamels on metal in the experimental phase at Guangzhou 廣州 in the second decade or two of the eighteenth century. However, nearly all later work from Guangzhou and Beijing was enamelled inside, for reasons to be discussed later.

The formalized floral design on the snuff bottle in figure I.16 is typically imperial. (The court being the centre of snuff culture at this time, even if this bottle had been made in a private workshop, the design would have still been dictated by imperial taste.) The bottle in figure I.17 has a feature that we find on a number of early enamelled metal wares from the Qing, the use of raised metal to separate larger areas of the design. In both bottles, the original gilding is all but lost entirely, suggesting extensive use and considerable age.

There are also several known early examples of better quality that feature only very limited use of the new enamels, usually just a dusty pink. If you were going to pick just one colour to use on cloisonné wares from the expensive new palette, you would chose pink as the most significantly different. Previously, an attempt at pink in cloisonné had resulted in a coarse mixture of red and white granules, whereas the new pink was homogenous and much prettier. The limited use of the new palette on the snuff bottle in figure I.18 suggests a date from the late Kangxi or the first five years of the Yongzheng era, between 1723 and 1728, before the costly imported enamels were replicated domestically.

The bottle in figure I.18 shares certain features with the bottles in figures I.19 and I.20. None is enamelled inside. They all have a relatively wide mouth, which we recognize as a feature of early metal bottles, and they all have a gilt-metal foot and foot rim, which seems to have been an initial, simple manner of making cloisonné bottles – one that was revived as an option on occasional later bottles or at certain workshops. Finally, they all show varying degrees of restraint in the use of the new palette.

The bottle in figure I.18 is likely to date from the late Kangxi era or the early Yongzheng era. Pink enamel is used, but in very small quantities and only in the surrounds to the main panels. The form is a typical early-Qing courtly snuff-bottle shape, and the twin-vajra design in the main-side panels, which is derived from Tibetan Buddhism, was also a popular palace motif during the Kangxi, Yongzheng, and Qianlong reigns.

Figure I.19 shows a bottle that uses a similar dusty pink colour, but more extensively. The ground colour, an unusual pastel turquoise green, may owe something to the new palette, too, but what is distinctive about it is that it has broken up with age, with overall crackling that looks like crazy paving under close examination. This would tend to indicate considerable age and a technology not yet under the perfect control, characteristic of the Yongzheng era.

Figure I.20 shows the famous example from the J&J collection, one of the most exciting cloisonné bottles known: the design is lovely, the date is early, and the matching stopper is original. There is a painted-enamel-on-metal bottle in the Denis Low collection, formerly in the Meriem collection, that has a matching original stopper in painted enamels and can be confidently dated to the last fifteen years of the Kangxi reign on the basis of its Kangxi yuzhi 康熙御製 (Made by imperial command of the Kangxi emperor) mark, so we know that such matching stoppers were an early feature, although they certainly continued into the later Qing.

The bottle in figure I.20 has an even less obvious use of the new enamels, with only some petal ends on the narrow sides possibly pink (rather than mixed red and white, which are used on the main-side lotus flowers and entirely in the traditional manner). But it does have two shades of pastel green, neither quite like that of figure I.19, but equally distinctive.

The lotus design seen here is a combination of arabesque flower designs originating in Syria and the aquatic lotus of India. The leaves in such designs are thin and twisting; the flowers are sometimes smaller than the aquatic lotus, sometimes large and spectacularly wild. Chinese descriptions of these designs refer to them as xifanlian wen 西番(蕃)蓮紋, (‘patterns of lotus from the western border regions’), because the design entered China through Central Asia. The term Indian lotus is often used in English for this design, although the term is also used for aquatic lotus of the subcontinent.

It is worth noting that, although several Kangxi-marked imperial cloisonné vessels other than snuff bottles exist, as do Qianlong-marked examples of the same sort probably numbering in the hundreds, only one is known with a Yongzheng mark. This indicates either that the Yongzheng emperor didn’t have most of his cloisonné wares marked or that the art form didn’t appeal to him. In the parallel case of palace moulded gourds, where Kangxi- and Qianlong-marked examples abound and there are no Yongzheng-marked ones, we are inclined to conclude that the Yongzheng emperor was uninterested in moulded gourds during his thirteen-year reign. Even though the records do suggest the existence of gourd snuff bottles, they must have been left unmarked.

A perennial problem in interpreting the records for details of cloisonné enamelling at court is that there is often some confusion between different types of enamelling, which include inlaid enamels such as cloisonné, champlevé, and repoussé,and painted enamels on glass, porcelain, or metal bodies. Some early records mention enamel workshops without saying which type of enamel was produced there. In the case of the painted enamels, it is typical not to mention whether the body is glass, porcelain, or metal. One early reference to gifts of enamels to a prominent figure from the emperor has been interpreted by one writer on enamelled metal ware as being painted enamels on metal, and by another, whose focus was enamelled glass, as being enamelled on glass. The Chinese maxim Renzhe jian ren, zhizhe jian zhi 仁者見仁,智者見智(The humane person [looks at someone and] sees humanness; the wise person [looks at the same person and] sees wisdom) would seem to apply here, except that both points of view cannot be correct.

Inlaid enamel snuff bottles unambiguously identified as cloisonné rather than some other form of inlaid enamel are absent from the Yongzheng records, although a small number may in fact be cloisonné. This is not surprising, if our supposition that the Yongzheng emperor was unenthusiastic about cloisonné is correct. The evidence for early painted enamels on both metal and glass is easier to find. Products of this sort from the Kangxi workshops are known: we can identify thirteen extant enamel-on-metal snuff bottles; surviving enamels on glass include a few vessels and one or two snuff bottles. Two of the glass pieces other than snuff bottles bear reign marks, allowing us to date them to the last decade or so of the Kangxi era.

We can be reasonably certain that porcelain snuff bottles made as such (not as a conversion) are not a Kangxi phenomenon. The earliest known porcelain snuff bottle to date is from the Yongzheng period. (There are a few pieces decorated in underglaze pigments and bearing Kangxi marks, but although they are sometimes published as being from the Kangxi era, they are very obviously from the nineteenth century.)

It is inconceivable that during nearly half a century of rapidly growing snuff addiction and demand for snuff bottles at court the Kangxi court would have overlooked nephrite, quartzes, soapstone, limestone, coral, mother-of-pearl, organic materials, metal, and mixed media as material suitable for the fashioning of snuff bottles in the palace workshops or elsewhere. Lists of types produced in the early Yongzheng confirm a wide range of materials pressed into service to make snuff bottles, including clamshells, tortoiseshell, ivory, coral, various stones, amber, and gourd. There is also a range of embellished bottles. Surely all or most of these had already made their appearance as snuff-bottle material in the late Kangxi period.

Now, it may well be that, because the best snuff was imported, snuff bottles made with imported techniques of glassmaking and enamelling were considered by the cosmopolitan-minded ruling elite to be most appropriate for holding snuff. And perhaps it was only after those technologies had been mastered for a decade or two that they lost some of their cachet and the attention of those who made bottles for the court turned increasingly to materials that resonated deeply within Chinese culture. Still, we find it difficult to believe that many other popular materials were not pressed immediately into service the moment the snuff bottle was invented as a separate category. Kangxi snuff-bottle craftsmen must have made use of such materials as nephrite, which epitomized beauty and even moral virtue in the Chinese mind throughout history; crystal, with its purity; agate, with its translucence, magical patterns, and hardness; coral, symbolic of longevity; amber and jet, with their associations with eternity; gourds, which we know were grown in the palace gardens and were a passion of the Kangxi emperor; and ivory. We can be sure that the expansion in the variety of materials used for snuff bottles did not begin magically when the Yongzheng era commenced in 1723. Most, if not all, of what we find in the early Yongzheng records must have been in place in the closing several years of the Kangxi era, if not earlier. There are, for example, so many post-1723 references to embellished and inlaid snuff bottles (sometimes involving repairs to bottles that must have been around for a few years), that the art had to have developed earlier—and the very fact that so few actual examples are identifiable today reminds us that 1) much is lost, and 2) much that is not lost cannot be recognized for what it is, given the limitations of our present knowledge. We must have at least a smattering of Kangxi examples in this range of materials in our collections, but which ones?

Given nephrite’s importance in Chinese culture and arts, we may speculate that snuff-bottle makers would have considered using it for bottles very early on. Figure I.21 shows a snuff bottle that is very much in the spirit of many small pre-Ming nephrite treasures, although the subject matter is typically Qing: gourd and butterfly, a rebus for perpetuation of the family. Its unusual extent of surface wear suggests a great deal of age, and supporting an early date is the design of the butterfly, whose distinctive willowy, exaggerated wings and long, calligraphically curved antennae are in accord with early Qing depictions painted on porcelain for the court. This may be one of the stone bottles we can date to the very first phase of hardstone carving in snuff bottles, a phase that lasted from the Kangxi period through into the early Qianlong. However, surface wear on a bottle like this is a weak indicator of age, for the skin of nephrite is far softer than the inner stone and wears more easily. It is true that the subject matter is a staple of Qianlong production, and the narrow mouth might indicate a Qianlong date as well, but we are still working with so little reliable data that we cannot be sure such features did not begin earlier.

There is much we have yet to discover about the state of the snuff-bottle arts in the Kangxi era, but we can say that by the time the Yongzheng era began in 1723, snuffing had taken firm hold at court and was spreading beyond the walls of the Forbidden City and the Imperial City to the rest of Beijing and surrounding areas. Officials going out on assignment or retiring to their homes elsewhere must surely have begun the process of spreading news of the addictive powder around the empire, as well. However, there is no indication that this prompted widespread production of snuff bottles for a non-northern, non-imperial clientele at so early a date.