It is known in the drinks industry that the visual appearance of a drink can strongly affect the attractiveness of the drink to the consumer.

Furthermore, adulteration, mislabelling and other deceptions practiced on the consumer or purchaser are known problems in the drinks industry. To counter these it would be desirable to be able to authenticate relatively quickly and cheaply a drink as coming from a particular source or supplier.

Consequently, in general terms the present invention provides a drink containing a fluorescent agent.

The fluorescent agent can provide the drink with a striking visual appearance which may be appealing to consumers. For example, the fluorescent agent can be selected such that the drink illuminates under UV lighting of the type commonly used in bars and nightclubs. Such establishments generally have low levels of ambient visible (white) light, and the fluorescent agent may be present in the drink in an amount such that fluorescence produced by the agent is visible unaided to the human eye under low levels of ambient visible light.

Thus in one aspect of the present invention, the fluorescent agent is present in the drink in an amount such that, in the presence of near UV light, the fluorescence is visible unaided to the human eye when the

ambient visible light is below 0.05 Wm-2 (-5 lux or-0. 1 pmols m-2 s-I).

Alternatively, the fluorescent agent may be present in an amount such that fluorescence produced by the agent is detectable but any non-fluorescent colour produced by the agent in the drink is not visible to the human eye. In this way, the fluorescent agent can serve as a "fingerprint"for the drink. The agent facilitates relatively easy authentication of the drink while remaining undetectable to the casual observer.

In a further aspect, the fluorescent agent comprises one or more phycobiliproteins. Phycobiliproteins are water soluble fluorescent proteins derived from cyanobacteria and eukaryotic algae. They are classified on the basis of their colour into two large groups, the phycoerythrins (red) and the phycocyanins (blue). Absorption maxima for phycoerythrins typically lie in the range from 490 to 570 nm while absorption maxima for phycocyanins are typically found in the range from 610 to 665 nm. These large groups have been further subdivided to reflect variation among the proteins in determining the exact location of the absorbance maximum and the specific shape of the absorbance spectrum. For drink additive applications, phycobiliproteins have, in general, the advantage of being tasteless and odourless.

Originally, these subdivisions, identified by letter prefixes to the phycobiliprotein, indicated the taxa of the organisms from which the pigments were isolated; e. g.

R-phycoerythrin was first isolated from the Rhodophyta.

It is now known, however, that specific phycobiliprotein types are not always restricted to specific taxa. They can be produced by different organisms or, on occasion, by

the same organism under different growth conditions, or at different stages of its life cycle. Thus, as used herein, the letter prefixes provide an indication of the shape of the absorbance curve, and do not limit the particular phycobiliproteins to specific taxa. In some pigments, isolated more recently, the name is followed by a number (e. g. phycoerythrin 566) which indicates the approximate absorption maximum.

Preferably the fluorescent agent comprises R- phycoerythrin, B-phycoerythrin, Y-phycoerythrin, C- phycocyanin, R-phycocyanin, allophycocyanin, allophycocyanin B, phycoerythrin 566, phycoerythrocyanin, phycourobilin, cryptoviolin, bilin 697, or a mixture of any two or more thereof. Advantageously many of these compounds, in particular C-phycocyanin and R- phycoerythrin, are obtainable from organisms which are generally recognised as being safe for human consumption.

Thus more preferably the fluorescent agent comprises C- phycocyanin and R-phycoerythrin or a mixture thereof.

In one embodiment the drink is an alcoholic drink e. g. containing at least 3% v/v (preferably at least 5,10 or 20% v/v) of ethanol. The drink may contain up to 50% v/v (preferably up to 40 or 30% v/v) of ethanol. Conventional alcoholic drinks generally have a pH in the range 6-8, while conventional soft drinks can have a pH of up to 3.

A reason for this difference is that alcoholic drinks rely to a greater or lesser extent on the alcohol to maintain sterility, whereas a low pH is often used to maintain the sterility of sugar-based soft drinks. Most phycobiliproteins (including C-phycocyanin and R- phycoerythrin) are stable in a pH range of about 4.5-9, which advantageously makes them suitable additives for alcoholic drinks.

For a strong non-fluorescent drink colouration and a level of fluorescence intended to be readily visible to the human eye, the C-phycocyanin, R-phycoerythrin or the mixture thereof may be present in an amount according to the respective formula C I 20, R > 10, or C/2 + R 2 10 (preferably C : 40, R 2 20, or C/2 + R > 20), where C is the concentration of C-phycocyanin in the drink in ug/ml and R is the concentration of R-phycoerythrin in the drink in ug/ml.

On the other hand, for"fingerprinting"purposes the C- phycocyanin, R-phycoerythrin or the mixture thereof may be present in an amount according to the respective formula C < 0. 5, R < 0.25, or C/2 + R < 0.25 (preferably C < 0.02, R < 0.01, or C/2 + Ru 0.01). With suitable detection equipment the fluorescence can be detected at very low levels of fluorescent agent. However, desirably at least 0.001 ug/ml of the C-phycocyanin, R-phycoerythrin or the mixture thereof is present in the drink, in order to provide at least a minimum level of fluorescence such that the use of e. g. a photomultiplier to detect the fluorescence can be avoided.

In a further aspect, the present invention provides for the use, in a drink, of a fluorescent agent comprising one or more phycobiliproteins. Preferred and optional features of the drink and the phycobiliproteins are as described in the previous aspects of the invention.

Another aspect of the present invention provides a method for authentication of a drink, wherein the drink, if authentic, contains a fluorescent agent, the method comprising the steps of:

(a) analysing the drink to detect fluorescence produced by the fluorescent agent; (b) determining the authenticity of the drink on the basis of whether fluorescence produced by the fluorescent agent is detected in step (a).

The fluorescent agent may comprise one or more phycobiliproteins. Preferred and optional features of the drink and the phycobiliproteins are as described in the previous aspects of the invention.

We have performed studies to test the feasibility of adding fluourescent agents to drinks. For these studies we used the phycobiliproteins C-phycocyanin and R- phycoerythrin.

The C-phycocyanin used in the studies was extracted from Anabaena cylindrica and from commercially available Spirulina using standard biochemical extraction techniques (see e. g. Oi et al., Journal of Biology, 93,981-986, (1982) ). For commercial production, C-phycocyanin extracted from Spirulina, which has a long history of human consumption and is FDA approved, is preferred.

The R-phycoerythrin used in the studies was extracted from Rhodymenia palmata using standard biochemical techniques (see e. g. Barrett and Bogorad, Biochemistry 10,3625-3634, (1971) ). Rhodymenia palmata also has a history of human consumption and is listed on the FDA's GRAS (Generally Recognised As Safe) list as a food additive which is generally recognised as safe. Rhodymenia palmata is also harvested and supplied commercially.

Each of the C-phycocyanin and R-phycoerythrin extracts was added to non-alcoholic and alcoholic drinks. Alcoholic

drinks containing up to 40% v/v of ethanol were tested.

The presence of ethanol at up to this amount did not have a noticeable effect on the fluorescent behaviour of the C- phycocyanin and R-phycoerythrin extracts. Furthermore, the C-phycocyanin and R-phycoerythrin remained in solution in the alcoholic drinks, i. e. they did not precipitate as many proteins do when exposed to ethanol. We expect these and other phycobiliproteins to form stable solutions in alcoholic drinks containing up to 50% v/v of ethanol.

The C-phycocyanin coloured the drinks blue and the R- phycoerythrin coloured the drinks and pink or orange/red.

However, when illuminated with broad-band visible (white) light (tungsten filament, or fluorescent tube), near-UV light (360-400nm), or UVA (320-360nm) C-phycocyanin produced an intense red coloured fluorescence with an emission peak at 647 nm and R-phycoerythrin produced an intense yellow coloured fluorescence with an emission peak at 575 nm. The fluorescence was, of course, most visible when the ambient visible light was at a low level. For example, the fluorescence was visible when the ambient visible (white) light was below 0.05 Wm-2 (-5 lux or-0. 1 umols m~2s~1) in the presence of near UV light. Such a combination of light conditions is often to be found in night-clubs and bars.

At a concentration of C-phycocyanin of 20 pg/ml in the drinks, red coloured fluorescence was readily visible to the human eye and also a strong, non-fluorescent blue colour was produced in the drinks. Higher concentrations (e. g. 40 ug/ml) produced more intense fluorescent and non- fluorescent colourations.

Similarly, at a concentration of R-phycoerythrin of 10 ug/ml in the drinks, yellow coloured fluorescence was

readily visible to the human eye and a strong, non- fluorescent pink or orange/red colour was produced in the drinks. Again, higher concentrations (e. g. 20 ug/ml) produced more intense fluorescent and non-fluorescent colourations.

Best results were obtained with clear drinks, such as those with a vodka base.

For"fingerprinting"drinks we found that under normal visible light conditions the non-fluorescent colourations produced by concentrations of 0. 5 ug/ml and 0.25 ug/ml of respectively C-phycocyanin and R-phycoerythrin were not detectable by the human eye. Fluorescence could be detected, however, at these concentrations by the human eye if the drinks were illuminated in the dark by a near UV, or UVA light source.

Fluorescence produced by significantly lower concentrations (0.02 pg/ml for C-phycocyanin and 0.01 ug/ml for R-phycoerythrin) was still detectable when a suitable long-wave pass filter was placed between the observer and the UV-illuminated sample (which was otherwise in darkness). Such a filter removes near UV or UVA light, thus allowing the faint fluorescence from C- phycocyanin and/or R-phycoerythrin present in the sample to be seen more easily. We found a 550 nm long wave pass filter to be suitable for detecting R-phycoerythrin fluorescence and a 600 nm long wave pass filter to be suitable for detecting C-phycocyanin fluorescence.