Introduction

Provision of adequate lighting to captive reptiles is undoubtedly one of the most important aspects of their husbandry. The full spectrum of natural sunlight encompasses the infra-red (perceived as warmth), "visible" light (as seen by humans) and ultraviolet light, subdivided into longer wavelength UVA, known to be visible to many animals, and shorter wavelength UVB which some reptiles appear to be able to detect, even if it is not actually visible to them.15

The provision of ultraviolet light to reptiles in zoos has received much attention in recent years. In view of the benefits of both UVA and UVB, most authorities now recommend that diurnal lizards, at least, be provided with appropriate UV light, either from natural unfiltered sunlight (in climates where this is possible) or from artificial sources, more of which become available every year. 1,2,54,55 Fierce debate continues as to whether it is advisable to provide all species, including snakes and crepuscular lizards, with UV light. Unsuitably high levels are likely to be detrimental; the requirements of species which do not bask in full sun in their native habitat needs careful evaluation 1 and in fact, this is already being carried out in a number of zoos today.39

Until relatively recently, few artificial sources provided anything but very low intensity UVB at short range.56 Specialist fluorescent tubes marketed for reptiles, described as 5% or 8% UVB, for example, when new typically emit only 15 – 25 uW/cm² at 12 inches. Whereas this may be ideal for reptiles with low UVB requirements (such as some species of chameleons in small enclosures), the use of these tubes to provide UVB to large lizards in roomsized enclosures is impractical to say the least.

Over the last two years, high-output mercury vapour lamps designed for reptiles have become available to the hobbyist and zoo keeper alike. Some brands, such as the lamps made by Westron of Canada, sold under the brand names TRex Active UV Heat Flood and ReptileUV Mega-Ray, now emit UVB at levels much closer to sunlight (150 – 250 uW/cm2) at distances of 12 – 24 inches.
The spectrum produced by a mercury vapour lamp, with emissions at discrete wavelengths, is very unlike the continuous spectrum of the sun. However, studies to date do not suggest that this adversely affects the production of vitamin D3 in the skin. Rather the reverse; there is growing anecdotal evidence 57 and at least two published studies 3,9 suggesting that such lamps do promote D3 synthesis and, when set up in suitably sized enclosures, may be beneficial to diurnal lizards which normally bask in full sunlight in the wild.
Spectrographic analysis shows clearly that these lamps do emit significant amounts of radiation within the D-UV range. Even these, however, do not produce a beam that will project high levels of UVB further than 3 to 4 feet into an enclosure.

For many years, some zookeepers and other herpetoculturalists in the UK have made judicious use of a human “tanning lamp”, the Osram UltraVitalux, which has a much higher UVA and UVB output than any of these lamps designed for reptiles.39
This lamp is not on sale as a reptile lamp to the general public in the UK, but appears to be in fairly widespread use by European herp keepers; it is available from several online stores for reptile hobbyists in Germany, for example. Only one version is available; a 300 watt, 230 volt internally ballasted lamp.

Another very high output lamp has recently appeared, however, this time marketed specifically as a lamp for reptiles in zoos or similar institutions, in very large enclosures where UVB is required at a considerable distance from the lamp. This is a lamp manufactured by Westron of Canada, the ReptileUV Zoologist Mega-Ray or “Zoo Mega-Ray”. It is available as either an internally ballasted 100 watt lamp, (120 volt or 220 volt) or as an externally ballasted 60 watt lamp with a separate ballast box, as a kit (120 volt only). The only difference between these versions is the way the lamp is ballasted.

This report describes both lamps in terms of their UVB output, decay (“burning in”), and beam characteristics.

Lamps on Test

For detailed testing, two lamps - one new Osram UltraVitalux 300 watt 230 volt internally ballasted Mercury Vapour Lamp, and one which had been in use for one year - were kindly donated by Douglas Sheriff, Specialist Keeper, Reptile Section, Chester Zoo, UK. Readings were also taken from another year-old Osram UltraVitalux lamp in situ in a vivarium at Chester Zoo. For detailed testing, one new ReptileUV Zoologist Mega-Ray 60 watt 120 volt externally ballasted Mercury Vapour Flood lamp kit was kindly donated by Robert MacCargar of ReptileUV (www.reptileuv.com).

Equipment

We use a UVB meter, a Vitamin D3 meter, a UVC meter, a lux meter and a UV-VIS spectrometer when testing UV lamps.

The UVB meter used for all tests is a hand held ultraviolet radiometer, the Solarmeter Model 6.2 UVB manufactured by Solartech Inc. (www.solarmeter.com) Although it measures the complete UVB range (280 nanometers to 320 nanometers) the Solarmeter's peak sensitivity is at 295nm, right in the middle of the D-UV range, which makes it useful for checking UVB reptile lamps.

The Vitamin D3 meter is a similar hand held ultraviolet radiometer, the Solarmeter Model 6.4 UVB manufactured by Solartech Inc. This device is designed to indicate the intensity of UVB radiation at the wavelengths that enable skin synthesis of vitamin D3 in human skin. The sensor's response follows the Diffey Erythemal Action Spectrum effective irradiance (Eeff), which closely follows the vitamin D action spectrum (Deff) in the short-wave ultraviolet (280 - 320nm).
We are currently evaluating the use of this meter in reptile husbandry, by taking readings from all reptile lamps, as well as solar readings.

The lux meter is a simple hand-held device, a SkyTronic LX101 model 600.620 digital lux meter (www.skytronic.co.uk) which measures the intensity of visible light, with maximum sensitivity around 560nm.

The UVC meter is the Solarmeter Model 8.0 UVC manufactured by Solartech Inc. This meter is only sensitive to ultraviolet emissions between 240 - 280nm. (UVC radiation)

The spectrometer we use is an Ocean Optics Inc. USB2000 spectral radiometer with a UV-B compatible fibre-optic sensor with cosine adaptor. (www.oceanoptics.com)

Testing Methods

The lamps on test were set up, in turn, in a ceramic holder affixed to a wooden batten close to the ceiling of the test room at a height of 90". This allowed measurement of UVB output up to 72" from the face of the lamp. For the final set of measurements taken over longer distances, the lamps were mounted horizontally on a batten affixed to a side wall, allowing measurement up to 11 feet from the lamp face.

Many UVB-emitting lamps demonstrate a rapid period of decay in UVB output (often described as "burning-in") over the first few days of use, followed by a slower decline in output over subsequent months. To establish whether this occurs in the lamp under test, and to ensure each lamp is tested in a comparable manner, a burn-in time of at least 80 - 100 hours is allowed for in all our testing regimes.
Most lamps take several minutes to reach full brightness. Tests are only performed on lamps which have been alight for at least 1 hour directly before testing (or 30 mins for the first test).

Five types of recordings were made for each lamp:

• direct readings
• spread charts
• spectrograms
• temperature recordings
• power consumption

Results

Because of the physical movement of the arc within the tube, and the way the radiation is generated, the UVB readings from mercury vapour arc tubes are rarely steady at close range, and in fact the visible light can often be seen to "dance" as well, casting flickering shadows at the edge of the beam. For consistency, the highest reading seen three times at each distance is the one recorded, as it is assumed that this reading represents the most accurate alignment of the hand-held meter with the beam.

The ReptileUV Zoo Mega-Ray EB lamp is a PAR-38 reflector lamp with a clear glass face with a moulded spreader lens on the surface. As with all externally ballasted mercury vapour lamps, when first switched on, the lamp produced a faint light from the arc tube which developed over a few minutes into a very bright light with (to the human eye) a purplish-blue tint and a directional beam.

The Osram Ultra-Vitalux is a larger ES/E27 lamp with an internal reflector and ground glass face. A self-ballasted lamp, when first switched on it emitted a very bright yellow-white light from the tungsten filament which changed in tone to a bluer tint as the arc tube lit up. The lamp illuminated a large area in a fairly uniform manner.

1. Direct Readings

(a) UVB recordings and Burning-in. A series of direct readings were taken with the UVB meter at increasing distances from directly beneath the centre of the lamp face. With new lamps, the first set was taken after 30 minutes burn; the next at 3 hours. The lamp was then allowed to burn for 15 hours a day and readings taken at intervals until a 90 hour burn-in had been completed. With old lamps, the readings were taken after one hour of use.

Figures 1a, b and c show the burn-in of the two lamps and their output at the end of the 90 hours.

Fig. 1a. UVB Output (uW/cm²) of ReptileUV Zoologist Mega-Ray during first 90 hours of use (Bar Chart).

Fig. 1b. UVB Output (uW/cm²) of Osram Ultra-Vitalux during first 90 hours of use (Bar Chart).

Successive readings taken over the initial burn of the Osram Ultra-Vitalux lamp showed a very marked "burn-in" decay taking place during the first few hours of use (Fig. 1a). For the first 15 hours all readings closer than 10" were off the scale (over 2000 uW/cm²) and at first, all readings closer than 22" were higher than normally seen in nature even at the equator (over 500 uW/cm²).
However, extremely rapid decay over the first few hours of use brought these readings down swiftly to lower levels. After 90 hours the reading at 18" was 420 uW/cm² - a value typical of hot tropical mid-day sunlight.
Nevertheless, all readings closer than this were still higher than normally seen in nature.

The rate of decay of UV lamps usually slows after the first few days, and this was indeed the case for this lamp. Between 45 and 90 hours burn, the precipitous fall in output seen earlier had become a slow decline with signs that it was levelling off. Over the entire burn-in period, however, the output dropped by about 34%.

In contrast, successive readings taken over the 90-hour initial burn of the ReptileUV Zoologist Mega-Ray lamp (Fig. 1b) did not show any marked "burn-in" decay taking place. The output only fell by about 2%. This lamp had a much higher output than the Osram Ultra-Vitalux from the beginning. Throughout the trial, all measurements closer than 26" exceeded 500 uW/cm².

Fig.1c. UVB Output (uW/cm²) of both lamps after first 90 hours of use (Line Graph).

The direct readings with the Solarmeter 6.2 meter show clearly that the output of both these lamps decreases exponentially as the distance increases from the lamp's surface. (Fig. 1c)

Figure 2 (below) shows the recordings taken from two older Osram Ultra-Vitalux lamps which had been in use in the vivarium for just over a year - a total of 3,300 hours (lamp ref. BO3) and 3,600 hours (lamp ref. BO4) along with the recordings from the new lamp when brand new and after 105 hours use.

Fig. 2: A Comparison of the Output of Three Osram Ultra-Vitalux 300watt lamps of different ages.

The much-reduced output of the older Osram lamps, as shown in Fig. 2, suggests that decay does continue, albeit at a much slower rate, as the lamps age. If we assume the old lamps had an initial output similar to the new lamp tested, they would have decayed about 65 - 68% in that year. Despite over a year's continuous use (over 3,000 hours) the output is still extremely high.
The year-old lamps still produce higher UVB levels than normal in nature, at all distances measured closer than 10" from the lamp. At 20" they are emitting about 140 uW/cm², which is still within the range seen for morning and evening sunlight in summer in the UK.

(b) Vitamin D3 and Lux Meter recordings. Sets of recordings were taken from each lamp using the Vitamin D3 meter and the lux meter. With the new lamps, these recordings were made after burning-in was completed.

Figure 3 gives the results for a set of recordings taken from each lamp, using the UVB meter (Solarmeter 6.2); the Vitamin D3 meter (Solarmeter 6.4) and the lux meter (SkyTronic LX101 model 600.620).
For comparison, direct solar recordings made on clear sunny days in Wales, UK at latitude 51n50 at several different times of day in June, September, December and March are included in the chart.
(Full details of solar recordings are available on-line in the UVB_Meter_Owners group files at: http://groups.yahoo.com/group/UVB_Meter_Owners/files/SolarUVB)

Solarmeter 6.4 Readings: International Units of Vitamin D3 per minute

The Solarmeter Model 6.4 is designed to indicate the intensity of UVB radiation at the wavelengths that enable skin synthesis of vitamin D3 in human skin. We are exploring its use in the evaluation of reptile lamps by comparing the readings from this meter with those obtained with the broad spectrum UVB meter and the spectral radiometer.

The units of measurement (I.U. of D3 per minute) do NOT indicate the amount of vitamin D3 produced by a reptile exposed to the lamp. The units are an estimate of the amount of D3 produced by a young adult human being with Caucasian skin type 2, with 10% of the skin (i.e., face, lower arms and hands) evenly exposed to that intensity of UV light, since the meter was designed for calculating human exposure to solar UVB, or UVB from sun-tanning lamps. Full details explaining the use of the 6.4 meter and its utility software are available on the manufacturer's website.
Reptile skin, however, is totally different from human skin in its permeability to UVB, and indeed, the skin of different reptile species varies markedly as well. (Some simple studies on this are found in our feature UV and Reptile Skin)

The sensor's response, however, closely follows the vitamin D action spectrum (Deff) and hence this meter should be less affected by the differing spectral power distributions of different lamps than the broadband UVB meter.
Using both the Solarmeter 6.2 and 6.4 together, it is also possible to assess the relative proportion of the UVB which is being emitted in the wavelengths mainly responsible for D3 synthesis (below about 305nm) within the total UVB range (280 - 320nm) and how the lamp compares with the sun in this respect.

These readings show clearly that both types of lamp not only emit high levels of UVB but also emit large amounts in the wavelengths responsible for D3 synthesis in the skin (the D-UV range).
The solar recordings show that the proportion of UVB in the D-UV range is much greater in the middle of the day, and in the summer; which is to be expected as shorter wavelength UVB is absorbed more strongly by the atmosphere. Hence the lower the sun in the sky, the greater the reduction in short-wavelength radiation. It is important to bear in mind that these recordings were made in the UK. At lower latitudes, at times when the sun is higher in the sky than is possible in the UK, not only will the total UVB be higher, but the proportion which is in the D-UV range will be significantly greater. At present the author has not been able to obtain comparable recordings from lower latitudes.

However, all 3 mercury vapour lamps tested emitted a higher proportion of their total UVB in the D-UV range than the sun as recorded from the UK. For example, the total solar UVB recording for June at 1pm in Wales, UK was 350 uW/cm² when the Solarmeter 6.4 reading was 61 IU/min D3.
The ReptileUV Zoo Mega-Ray was emitting a similar total UVB reading at 32" distance (322 uW/cm²) but at that distance, 148 IU/min D3 was recorded. This lamp, at that distance, would appear to be over twice as effective at producing D3 as the sunlight was on that June afternoon.
Likewise the new Osram Ultra-Vitalux was emitting 368 uW/cm² total UVB at 20" where 189 IU/min D3 was recorded. The older Osram lamp had a much-reduced UVB output. 363 uW/cm² total UVB was emitted at 12" but at that distance only 133 IU/min D3 was recorded. Although this is still more than twice as effective as the sunlight in our example, it would appear that the decay has affected the lower wavelengths of the lamp slightly more than the higher ones.

From a practical point of view, these readings indicate that at the centre of the beam, a mercury vapour lamp is in fact emitting a higher level of UVB light in the D-UV range than the overall UVB reading might suggest.
Readings comparable with the sunlight at 1pm on that June afternoon would be achieved at distances of slightly less than 20" with the old Osram lamp (57 IU/min D3), 36" with the new Osram lamp (64 IU/min D3) and 48" with the ReptileUV lamp (65 IU/min D3).

Lux meter recordings

Typical published values for a brightly lit office, and for sunrise or sunset on a clear day, are around 400 - 500 lux. Those for a clear spring morning, 30 minutes after sunrise in the UK, are around 10,000 lux.58,59 These are "global" readings with the lux meter aimed at the ceiling or zenith, not direct recordings from the light source or the sun. Our recordings complement these figures, with (for example) a global reading of 6,000 lux and a direct solar reading of 31,600 lux at 5.00am GMT with clear skies on 21st June 2005. At 12.30pm GMT on 26th June 2005, also with clear skies, a global reading of 117,500 lux and a direct solar reading of 134,300 lux were taken.
The brilliance of full sunlight on a summer's day cannot be matched by these mercury vapour lamps at distances which would be suitable for a basking lamp. Direct recordings made underneath the lamps on test reveal that at a distance of 3ft, these illuminate a basking spot with about 6 - 8,000 lux; at 4ft the light is between 3 - 5,000 lux. Although this is far superior to anything provided by most UVB fluorescent tubes we have tested (range: 350 - 630 lux at 1ft) it is still, at best, only providing illumination, in terms of "visible light", similar to that of the sun around dawn and sunset.

(c) UVC Readings

Figure 4 gives the results for recordings taken at close range from each lamp with the UVC meter (Solarmeter 8.0).

All three lamps emit very small amounts of light in the wavelengths between 240 - 280nm. No lamp produced any measurable UVC at 12" distance.
Although UVC is harmful to living cells and ideally, no lamp should emit UVC, most lamps which utilise mercury vapour do emit traces at extremely close range. All fluorescent tubes we have tested, for example, give readings of between 1-3 uW/cm² of UVC at the tube surface; some occasionally register 1 uW/cm² of UVC at 1" distance. The closest we were able to test mercury vapour lamps (because of their high heat output) was 1-2"; all except for one brand we tested produced at least 1 uW/cm² at 1" and all except for two brands produced between 1 and 36 uW/cm² at 2".
However, at the distances at which the ReptileUV Zoo Mega-Ray and the Osram Ultra-Vitalux are suitable for use in the vivarium, the UVC component is not of concern. With these lamps it is undetectable with our meter (i.e. below 1 uW/cm²) at 12". UVC radiation only travels very short distances through air and it seems unlikely that any would reach a basking spot several feet below one of these lamps.

(d) Long Distance Measurements

The lamps were mounted horizontally on a side wall for this set of readings. Recordings were taken from each lamp, at increasing distances from the lamp face, using the UVB meter and the Vitamin D3 meter, up to a maximum distance of 11 ft. The results are shown in Figure 5.

Both the ReptileUV and Osram lamps have a powerful enough output to ensure that UVB is projected a considerable distance from the lamp. The output from the new 60-watt ReptileUV lamp is, however, approximately twice that of the new 300 watt Osram lamp at all distances.
The desired level of UVB at a basking spot must be determined by consideration of the basking habits of the species in question and the ambient levels of UVB found in its natural environment. For example, a study was recently made of a wild female Cayman Island Iguana (Cyclura lewisi) 60. Moving in and out of shade, she kept herself exposed to about 100 uW/cm² all day. From 8.30am to 3.30pm she remained in locations where the UVB was between 74 -162 uW/cm². She fully exposed herself to light from bright but cloudy skies early and late in the day and moved in and out of lightly shaded areas in the middle of the day, when full sunshine with a peak UVB intensity of 447 uW/cm² was recorded.
The ReptileUV lamp would provide a UVB gradient up to approximately 100 uW/cm² if placed at a distance of 5ft above the basking spot. The newer Osram Ultra-Vitalux would provide this at a little over 3ft.
However, as we have seen, the proportion of UVB which is in the D-UV range, in a mercury vapour lamp, exceeds that found in sunlight. Lamps sited at these distances will provide, at the centre of the beam, higher UVB in the D-UV range. Without knowing the exact proportion of UVB which is in the D-UV range in daylight in the shade on Grand Cayman Island, we cannot calculate what the equivalent illumination from a mercury vapour lamp would be. We can make a rough estimate. Under lightly overcast skies at mid-day on June 21st 2005, here in Wales UK, a reading of 103 uW/cm² with the Solarmeter 6.2 was taken simultaneously with a Solarmeter 6.4 reading of 17 IU/min D3. If the proportion of UVB in the D-UV range is similar to this, in our example, then an equivalent distance for the ReptileUV lamp would be a little over 8ft and for the newer Osram lamp, it would be 6ft.

2. Spread Charts

Whether a lamp is useful in the vivarium does not merely depend upon the intensity of its output as measured directly under the lamp. The shape of the beam, and hence the area of the vivarium which is illuminated with UVB light at the required intensity, is just as important. Our tests on each type of lamp therefore included the construction of a spread chart; this is a useful way of visualising the three-dimensional beam, by plotting the UVB gradient within a cross section of the beam.
The output of the lamp is measured in a two-dimensional plane directly beneath and to the sides of the lamp face. Direct readings are taken from several hundred points in this plane, and plotted on a chart so that a two-dimensional visualisation of the three-dimensional "cone" of radiation emitted by the lamp can be created.

(Details of how such charts are made is described in the feature: Constructing a Spread Chart)

Figure 6 (below) is the UVB spread chart for the new Osram Ultra-Vitalux lamp (ref. BO1) after 109 hours burn, showing the contours plotted for UVB output from 10 uW/cm² upwards.
Figure 7 is the chart for the Osram Ultra-Vitalux which has been in use for one year (ref. BO4), plotted to the same scale.
Figure 8 is the chart for the ReptileUV Zoo Mega-Ray lamp after 80 hours burn, showing the contours plotted for UVB output from 20 uW/cm² upwards.

Osram Ultra-Vitalux lamps

The Osram lamps are flood lamps, producing an extremely wide beam of UVB radiation projected beneath the lamp. For most of its length, all radiation from 20 uW/cm² upward from the newly burned-in Osram lamp is contained in a roughly cylindrical zone over three feet wide, with a gradient towards its central axis.
At a distance of 2 feet, the beam (defined here as radiation from 20 uW/cm² upward, projecting from the lamp) is about 3.5 feet wide and reaches over 240 uW/cm² at its core. This is around the level recorded from direct mid-day sunlight in May, in the UK. At a distance of 3 feet, the beam is still 3 feet wide and the highest reading was just over 60 uW/cm2 in the centre here. The beam tapers gradually but at six feet from the lamp surface, 20uW/cm² (cited by one author as a minimum level recommended for green iguanas 27) is still available in a zone about 30 inches wide, with just over 30 uW/cm² being recorded at the centre of the beam.
The spread chart for the one-year-old Osram lamp shows, as might be expected, a scaled-down version of the UVB gradient seen with the new lamp. At a distance of 2 feet, the beam is not quite 2 feet wide and reaches nearly 100 uW/cm² at its core.

ReptileUV Zoo Mega-Ray lamp

This is described by the manufacturers as a narrow flood lamp, and indeed, the beam is not as wide as that of the Osram lamp but it is projected further beneath the lamp. The lamp produces a very long teardrop shaped beam which, close to the bulb, does not extend far to either side of the bulb, except for a curious "spur" effect at the top of the beam, which is caused by light bypassing the outer edges of the hexagonal moulded spreader lens. Visible light also emerges at these points and can be seen to make a regular pattern on adjacent walls.

Three feet below the lamp, the beam is about 28 inches across and reaches around 250 uW/cm² at the centre. (This is the manufacturer's recommended minimum basking distance.) Further away from the lamp, the beam widens. Six feet from the lamp, the beam produces a footprint approximately 30 inches across, which is similar to that of the new Osram lamp in coverage, but the UVB gradient rises much higher, to 70 uW/cm² at its centre.

3. Spectrograms

Spectrograms were recorded from each lamp under test, using an Ocean Optics Inc. USB2000 fibre optic spectrometer, model UV-VIS (178nm-850nm), with a UV-VIS fibre, and a cosine corrector. All recordings were taken at a standard distance of 30cm from the lamp surface.

Figure 9 shows the full UV and visible spectrum of a mercury vapour lamp, in comparison with that of the sun and of a halogen lamp, which, like other types of incandescent bulb, is frequently used as a basking lamp.
The absolute irradiance of these three is not shown owing to their extreme differences in magnitude; the graph shows only the relative spectral power distributions, to enable comparisons to be made.
Overlaid onto this image is the outline of the action spectrum for the conversion of 7-DHC to pre-vitamin D3; this gives an indication of the location, on the spectrogram, of the wavelengths of light that are effective in enabling the synthesis of vitamin D3 in the skin.

Figure 10 shows the recordings from the ReptileUV Zoo Mega-Ray and the two Osram Ultra-Vitalux lamps, in the UVB wavelength range (280 - 320nm) all taken at the same integration time (200msec) and at a standard distance from the lamp face (30cm).
The spectra are therefore comparable although the spectrometer was not calibrated to measure absolute irradiance (uW/cm²/nm) when this recording was made, hence the readings only indicate relative intensity.
Again, the outline of the action spectrum for the conversion of 7-DHC to pre-vitamin D3 is overlaid onto the image to indicate the wavelengths of light that are effective in enabling the synthesis of vitamin D3 in the skin.

The sun produces a continuous spectrum from UVB (around 290nm) to infrared, with a peak at about 460nm (Figure 9). The halogen lamp spectrum is a bell curve typical of "black body radiation" but it is lacking in ultraviolet and blue, and has much more of its output in the red wavelengths, peaking at around 650nm.
Externally ballasted mercury vapour lamps, such as the ReptileUV lamp shown in Figure 9, produce only spikes of radiation at the specific wavelengths typical of a high-pressure mercury vapour arc. Because there are large spikes only in the UVA, purple, blue, green and yellow wavelengths, the light from these lamps appears, to the human eye, blue-green and lacking in "warmth". (Hence the warmer tones of the halogen lamp complement them well.)
The self-ballasted lamps such as the Osram Ultra-Vitalux also produce visible light and heat from the tungsten filament. This does not alter the ultraviolet spectrum of the lamp in any way, but the spectrum from a tungsten lamp contains a higher proportion of yellow and red wavelengths, again "warming" the tone of the lamp.

The only wavelengths involved directly in vitamin D3 synthesis are those in the UVB and very low UVA range (280 - 335nm). The action spectrum for the conversion of the cholesterol 7-DHC to pre-vitamin D3, as described by MacLaughlin et al 34 is shown as an overlay in Figures 9 and 10. The wavelengths involved in this conversion are those below 315nm (the "D-UV range" described above). Wavelengths between 280nm and 335nm are also involved in several photochemical control mechanisms preventing excessive vitamin D3 production.

The amount of light produced at these wavelengths is very small indeed. Figure 10 shows the spectra of the three lamps on test, in this wavelength range. These lamps clearly emit measurable UVB in the range required for vitamin D3 synthesis. Mercury vapour lamps produce small spikes of radiation at 296nm and 302nm and a larger spike at 313nm. The former two are in very good positions for promoting D3 synthesis. The first, smaller peak is at the optimum wavelength for the photoconversion process, and the larger peak is still well within the action spectrum.

4. Temperature recordings

An infra-red non-contact thermometer, Electronic Temperature Instruments Ltd Model TN-1 (www.etiltd.co.uk) was used to record heat radiated from the lamp towards a "basking spot", and also the heat conducted through the ceramic lampholder to the surface on which the lamp was attached.

To assess the heat radiated from the lamp towards a "basking spot", a series of measurements were taken of the surface temperature of a piece of untreated light brown plywood on the test bench, with and without illumination from the lamp on test, which was mounted at distances of 2ft, 3ft and 4ft above the wood. To ensure the temperatures had completely stabilised, the lamp was set in each position for an hour and a half before each measurement was taken.
The results are shown in Figure 11.

When using any lamp in a vivarium, it is vital to measure the temperature directly under the lamp, at the basking spot, to be certain that the reptiles cannot receive thermal burns.
If these lamps are to be used in large enclosures, however, at suitable distances from the basking spot to maintain natural UVB levels, they are likely to be too far from the animals, in most cases, to be used as a primary heat source.
As can be seen fro