The subject of UVB lighting for reptiles can be confusing at times. A company in the UK tests the quality of a wide range of reptile bulbs available throughout the world. The following are excerpts from emails I've sent to Francis Baines on the subject. You can check out their web site on http://www.uvguide.co.uk
In the first email (sent in 2009) I asked about the sylvania brand of tubes that I purchase for my turtle rescue work. The answer is as follows:


The tubes that have proven hazardous, to date, have all been of Chinese manufacture. The Chinese use combinations of glass and phosphors which can allow hazardous, abnormally short-wavelength UVB and UVC through. I tested some Chinese samples a while back, which Hot House Turtles suspected could be dangerous. They were right.

However, Sylvania Reptistar tubes have a very good spectrum. They are ideal in my opinion for use with reptiles. They are made by a German company and are very high quality.
I tested one about a year ago and the results were very good. Here is the "Conclusion" from the report I wrote for that lamp - a 24", 18-watt T8 tube:

Conclusion
These results indicate that this lamp has a spectrum which is very similar to that of sunlight in the UVB and low UVA wavelengths and does not have any significant output at lower wavelengths which are not found in natural sunlight. This is close to the ideal spectral power distribution for the production of vitamin D3 in the skin. This lamp produces UVB in the wavelengths required for vitamin D3 synthesis, but the output is low and even at fairly close range (between 15 and 30cm) never exceeds UV Index 1.0 - that of a heavily overcast sky or relatively deep woodland shade.

This lamp is likely to be appropriate for general use in a small vivarium where it can be sited appropriately above the reptile’s basking area. It might be suitable for species which normally live in habitats shaded from intense sunlight. Species which, in the wild, bask in intense sunlight would normally require a higher UVB level at the basking spot, since very prolonged exposure to this lamp would be needed to produce the equivalent ‘dose’ of UVB radiation to that provided by short-term exposure to full tropical or sub-tropical sunlight.

The lamp does not produce significant amounts of visible light and the spectrum is discontinuous. In most situations, it would not provide sufficient visible light, on its own, to simulate daylight in a vivarium. Like all UVB fluorescent tubes, it would be best combined with an intense full visible-light spectrum lamp (such as a halogen or metal halide lamp) to produce a basking spot resembling natural sunlight.


You definitely want to remove ALL plastic or glass from between lamp and reptile. UVB (which they need to produce vitD3) is completely blocked by almost all ordinary glass and plastics. Some of the light fittings - even ones sold for reptiles!!! - have so-called "transparent" covers which say that they let all the beneficial rays through, but believe me, they don't!! I haven't found a single one which is any good.

Something else it's worth knowing, is that the UVB from a tube is strongest towards the middle of the tube and tapers off at the ends. So if you hang one long tube over several tanks, the end tanks will go short..
I don't have a diagram for a Reptistar but the ZooMed Reptisun 5 tube is almost exactly the same. (Actually I believe from looking at the spectrum, that they are the same.... they both come from Germany....) Here is what the beam looks like: (I made these diagrams for a talk I did for the British Chelonia Group last year)

Image

Sylvania Reptistar tubes (like all fluorescent tubes) do not produce the same amount of UVB as direct sunlight. As you can see, they produce about the same amount of UVB as you'd get in outdoor shade on a sunny day, if you hang them at about 8 -10 inches above the turtle's back. You can improve this by fitting an aluminium reflector strip behind the tube, so it reflects all the light, including the UV, down towards the turtle... you can almost double the UVB... but although this is better (and I'd recommend doing it), it's still not like real sunshine.
If you have species like some tropical Asian turtles that live in leaf litter or mud, and in forest shade, these tubes would be ideal.

If you have species that bask in full morning sun, like the American Red-eared-slider turtles so popular in the pet trade here - and Mediterranean or African tortoises that live in dry arid grassland and would get a great deal of sunlight - especially if they are suffering from metabolic bone disorder from insufficient vitamin D3... I would recommend one of two things.

First - can you give them real sunshine? Not filtered through glass or plastic of any sort? Outdoor tortoises and terrapins usually get plenty of UVB all summer in the UK, and NZ is mostly at more sunny latitudes than here...
Secondly, if they must be indoors - do consider high quality mercury vapour lamps, if you can get some. If only for treating the worst cases of MBD. I don't know what you can get in New Zealand, but I would suggest avoiding anything made in China, and anything with a totally clear glass face. Some of them have dangerous pencil-thin beams of UVB that are far too high. Others produce little UVB at all...I've tested a couple of Australian mercury vapour lamps.

The "Oz Bright" lamp from UltimateReptiles.com.au was very poor quality. It was Chinese-made. The UVB beam was little better than from a fluorescent tube.

The ReptileUV MegaRay from AnimalAttraction.com.au is a Canadian lamp made for a small but dedicated home-grown company in the USA, and that lamp has always performed best in all the tests on UVB mercury vapour lamps I've ever done. Most turtles that bask in sunlight would do well under a 100W MegaRay lamp hung 12 - 14" above them when basking. With these you get UV Index 5 - 7 (ie. strong mid-morning sun) at the centre of the beam. But you need quite good ventilation and reasonably big tanks because all mercury vapour lamps give off heat as well as light and UVB. Here's what the beam looks like. This was a 60W EB lamp (needs separate ballast box) but the beam is the same with the SB (self-ballasted) lamps.

Image

This year I queried the use of tube versus compact lighting and also a few different brands. While there is alot of technical stuff, it does make very interesting reading and explain some excellent guidelines on the safe use of bulbs with reptiles. The email follows:


These are the main points from your emails (to make sure I answer all your questions, I'll paste them here...)
I have been involved in a discussion on a forum thread in the States (Happy Turtles Pub), that advocates using a Reptisun 10% UVB 40watt fluoro bulb for the RES and other turtles and this is being questioned due to those levels causing skin cancer in humans. Here in NZ we can really only access 6% or 10% UVB reptile bulbs and its being proposed that 4% is safer?
For the Eastern Long Neck Turtles, the Australian web site (Australian freshwater turtles), advocates using a NEC T10 Blacklight Fleurescent tube, only about 10cm away from the basking area. This seems quite close to me? Could there be damage to the turtles eyes and is this a good bulb for turtles?
I've never known if the 'tube' light bulbs are better then the 'compact' reptile bulbs either? Something people ask me about as well.
The common brand here is the Repti glo 5.0 or 10 UVB by Exo Terra. The other is the Sylvania Reptistar T8 18W or 15W which is cheaper and available wholesale. I know you gave me info on that one in the past. I believe its not so good to use on the RES as it is better suited to a tortoise in a wooded area then the RES who needs full sunlight. I've passed the info onto the import company but I guess they wouldnt know which brand was really suitable as they are a general wholesale lighting company. I figure its better that people use the cheaper brand with less UVB and change it 6mthly then not change their bulb at all? So many people use the original bulb they bought at the pet shop and 2,3, 6 yrs later.... the turtle ends up here with a shell like a squishy meat pie :(
Here's some other types of bulbs being used for turts by people in the States and Portugal that people are wondering about their effectiveness for aquatic turtles. ExoTerra ReptiGlo 26W 10.0 UVB Desert Terrarium lamp and ExoTerra Repti Glo 10.0/T8 Fluorescent tube (various lengths)


OK, I'll start off with a brief description of the types of fluorescent lamp available and what the numbers and wattages, etc., actually mean. It's quite confusing. Then I'll explain the difference between the brands.

T8 fluorescent tubes are the standard, long tubes. They are 1" thick (25mm). The T8 refers to the diameter of the tube in eighths of an inch (yes, archaic!), so a T5 is 5/8ths of an inch and a T8 is 8/8ths of an inch - i.e., 1 inch! and the T10, rarely used nowadays, but used by NEC for their Blacklight, is 1¼ inches across.
The wattage has nothing to do with the intensity of the UVB. It just reflects the length of the tube, e.g., an 18" (45cm) T8 tube is normally 15 watt; a 24" (60cm) tube is 18 or 20 watt; a 30" (75cm) tube is 25 watt and a 48" (120cm) tube is 40 watt. You need more electricity to fire up a longer tube. Of course, you get a lot more UVB from a 120cm tube than a 60cm tube, but that's because there's a bigger surface area emitting UVB, not because the lamp emits more UVB per square cm. of its surface. In other words, it lights up a lot more of the tank because it covers more of the tank, not because it has more "zap". The turtles can bask at a sensible distance (say 20 - 25cm) away from any point along a good brand of tube; a long tube can make a large basking zone.

However, if you make a fluorescent tube thinner, such as used in a T5 tube, or a compact lamp (which is rather like a thin, coiled-up T5 tube, you might call it a T2.5 or T3, I suppose!) the mercury vapour arc is closer to the walls of the tube. The phosphor coating is heated to a higher temperature, which increases its output. Also, the UV light produced by that lamp, which is related to the power consumption (ie., wattage) is being emitted over a smaller surface area than a T8. So these narrow tubes burn hotter, brighter and with more UVB per square cm. of their surface than a conventional T8 tube.
In a compact lamp, all the UVB-emitting glass is squashed into a relatively small area, though. So these are very hot, intense lamps... not very pleasant at all to look at - but all the UVB is produced in a tight zone very close to the bulb. At close range, the UVB from some brands of compact lamps can be much too high and there is a powerful glare from the UVA and blue visible light, too... hence, compact lamps are more likely to cause eye problems, at close range, than T8 tubes.
But this intensity is only found at very close range. Unlike torches and car headlamps, and the "reflector" bulbs used in basking lamps, these compact lamps don't make a focused beam. So the UVB is very strong, close to the lamp BUT it has a very steep gradient.... at a sensible basking distance, say 25-30cm, the basking zone from even the best lamp is very small. That's another reason why I personally don't like them for basking reptiles...
My final reason for generally tending to avoid compact lamps (except for use in very small vivaria in certain circumstances) is that if you put them in an aluminium reflector - either a dome or a reflector hood - you CAN create a focused beam. The aluminium catches up all the UVB from the whole surface area of the lamp and directs it strongly downwards... but the beam that's created depends upon all sorts of unknown factors like the exact length of the lamp and curvature of the dome, and the angle at which the lamp is held in the reflector.... Without a UVB meter, with some dome-lamp combinations, you could be creating something like a six-times increase (or more) in the UVB at a certain distance beneath the dome. I wouldn't risk this without checking the output at the basking zone with a meter, to make sure it isn't too high.

Next, we've got the usual problem of the 2%, 5%, 10% UVB to think about. The "percentage of UVB" simply means the proportion of UVB compared to visible light.
The divisions are arbitrary (scientists don't even agree about what should be called UVB, what should be called UVA and where Visible Light goes up to) but in the USA, UVB is UV light with a wavelength of between 280nm and 320nm; UVA is from 320nm to 400nm, and Visible Light is the "rainbow" of colours from 400nm (purple) to somewhere between 700 - 800nm (red). (I use 750nm as a cut-off in my work.) Above that, is infrared, usually called "heat".
When the first reptile UVB lamps were brought out, the manufacturers didn't know how to indicate the "strength" of the UVB, so they decided to express it in this way, because it doesn't vary with distance from the lamp, like measurements of UVB intensity do.
In theory, a 5% UVB, 30% UVA lamp (as some have been described) should have 5% of its output in the UVB range, 30% in the UVA range and 65% Visible Light.
The amount of heat emitted (infrared radiation) is ignored.
You can see the weakness in this method quite easily, because it relies upon proportions, not actual amounts of light emitted. So a lamp with a poor level of UVB but which has very dim visible light could have a higher % UVB than a strong UVB lamp with especially bright visible light. An unscrupulous lamp manufacturer could put the same amount of UVB-emitting phosphor in all his lamps, so they all emitted exactly the same intensity of UVB, but alter the percentage of UVB from 5% to 10% by simply decreasing the amount of visible light-emitting phosphor, thereby making the visible light 5% dimmer. Admittedly, this is probably rare... but I have seen it, once.

What I find more often is that the 5% or 10% or 8% or whatever, is fairly accurate for some brands but not at all accurate in others. Many manufacturers don't use the % sign any more because of this; they call the lamp 10.0 or 5.0 or 8.0 and so on. I think the best interpretation of these numbers is probably "Brand X's 10.0" is supposed to be stronger than "Brand X's 5.0". But whether it bears any relationship to Brand Y's 10.0 and 5.0 is anyone's guess....

I have been involved in a discussion on a forum thread in the States (Happy Turtles Pub), that advocates using a Reptisun 10% UVB 40watt fluoro bulb for the RES and other turtles and this is being questioned due to those levels causing skin cancer in humans. Here in NZ we can really only access 6% or 10% UVB reptile bulbs and its being proposed that 4% is safer?
I think this may have arisen because of a confusion over the way that the %UVB is described in human tanning lamps. These give the %UVB as a percentage of the Total UV only (ie, UVB+UVA) and NOT including the visible light. (Totally different. Don't ask me why!) Tanning lamps are almost entirely UVA, with a very small amount of UVB. So a so-called "high-UVB" tanning lamp would be described as 4% UVB and 96% UVA. Usually, I think 3% UVB is the limit allowed for humans, but I'm no expert...
Just out of interest, I worked out what a Reptisun 10.0 would be if it was measured as a human tanning lamp. It's 26%. This is because reptile lamps have very little UVA and a lot more UVB. It does not mean they are more likely to produce skin cancer than tanning lamps though... because the actual amount of UV they emit is a great deal less. I tested a 25W human tanning lamp, and its total UV output was five times greater than a 20W Reptisun 10.0.


So I think, now I've covered those points, I'll look at the individual brands you've asked about.
When considering a new lamp, two things are vital -
(1) The wavelengths of UVB that the lamp emits (which determines whether it can produce vitamin D, and also what the risks are for skin cancer and eye damage). This depends upon the phosphor blend used.
and
(2) the intensity of the UV, which is related to the wattage, but is also dependent upon the phosphor blend used.

1. Wavelengths.
The sun emits UVB and UVA as well as visible light, and that UVA and UVB can damage the skin (sunburn) and eyes (photo-keratitis and photo-conjunctivitis.) UVA damages the skin cells in a different way to UVB, but both can cause "burning". If the skin is badly burned, and some skin cells don't die but survive with damaged DNA, they very rarely go on to become cancerous.
Sunlight is vital for life, and the risk is actually quite low, but we can never say any UV is completely "safe". But sunlight is probably the safest way of getting effective amounts of vitamin D (for reptiles and for humans), because it has the completely natural balance of UVB and UVA that living things evolved to thrive under.
Here is a spectrum of typical summer sunlight, in the UV wavelengths that are involved with vitamin D production, so I can show you what I mean. It's the green line. This graph doesn't tell you the intensity of the light (ie., microwatts per square centimetre) it is "normalised" to show the percentage of maximum, in the range shown (270 - 350nm, the wavelengths most important in D3 synthesis and UV effects upon skin). This means we can compare any light sources whether very bright or very dim, and see how the "shape" of their spectra vary (the proper term is spectral power distribution)....

Image

Mid-day sunlight in summer across most of the higher latitudes (including New Zealand and the UK) has about 295 - 300nm as its shortest wavelength, and as you can see, there are only tiny amounts of this. (At the equator at mid-day, you can get tiny amounts of even shorter wavelengths, down to 290nm, but few reptiles go out in overhead sunlight). Then as the wavelengths get longer, the amount of each wavelength in the light rises; there is a more or less steady increase with increasing wavelength, up through the UVB and UVA. By the time you get to the boundary between UVA and purple visible light, at 400nm, there's 100 times more than at 295nm.
No lamps can match the exact proportions found in sunlight. But the less a lamp spectrum resembles sunlight, the more hazardous it has proven to be. A couple of years ago, several brands existed which had extremely short-wavelength UVB in their spectrum, starting at 275nm - 280nm instead of 295 - 300nm.

If you look at the graph, I've drawn two curved lines in red and black. These are called "action spectra" and they are a way of showing how effective a wavelength is, at having an "action" on something else. Let's look at the black one - the Pre-Vit D3 action spectrum. It starts up off the edge of the graph to the left - about 250nm - and is up to maximum at about 295-300nm, but comes down again and tails off at about 320nm. What this means, is that light at 295-300nm is the most effective at producing Pre-vitamin D3 (which is the first step in making vitamin D3). Longer wavelengths than this are progressively less efficient at it, and UVB at 315 - 320nm is very inefficient - you need a vast amount of it to make the same amount of D3 that a tiny flicker of UVB at 300nm will make. And light above 320nm doesn't make D3 at all.

So you might say, "sunlight is not very good at making vitamin D3. A lamp with more of its UVB around 295 - 300nm would be better."
But that's where the trouble begins, because shorter wavelengths don't just make vitamin D3. They start to damage DNA. The red curve shows an estimate for "DNA damage", and you can see that if you have a high proportion of UV below 300nm, you greatly increase the risk of DNA damage, and the shorter the wavelength, the worse it gets. Sunlight is the perfect UVB source; it maximises the vitamin D3 synthesis with minimum risk of DNA damage.

Now let's look at the spectra of the five fluorescent tubes that you've asked me about:
ZooMed Reptisun 10.0
Sylvania Reptistar
NEC T10 Blacklight
ExoTerra ReptiGlo 10.0
ExoTerra ReptiGlo 5.0

Here's the chart, with their spectra on it

Image

All of these spectra are similar to sunlight in general "shape" starting off with a small proportion of UVB in the shortest wavelengths, rising with increasing wavelength to about 340nm... except for the ExoTerra ReptiGlo 10.0, which has much less UVA. (You can see how the blue curve falls, after the peak at 313nm, whereas the others do not.) UVA is visible to reptiles, so this lamp will look less bright to them.
But notice where the "threshold wavelengths" are, ranging between 285nm and 305nm. This gives some idea of the relative "safety" of the spectrum, compared with sunlight, and also how effective it will be at making vitamin D3.

The only spectrum with a threshold wavelength longer than sunlight (-it has nothing below 305nm, and very little below 310nm -) is the NEC Blacklight. This lamp has the lowest proportion of its output in the wavelengths responsible for vitamin D3 synthesis. So to produce the same amount of vitamin D3 as sunlight, you will need a higher "dose". We might predict that it will need to be placed close to a reptile to have a good effect; but the absence of short wavelength UVB also makes it unlikely to be high-risk in terms of eye or skin damage, even at fairly close range.

All the others have more of their UVB in the shorter wavelengths than does mid-day summer sunlight in the UK; but there is a great deal of difference between the brands.
The Sylvania Reptistar has the spectrum closest to sunlight, in this range of wavelengths. We can predict that this lamp will not have a particularly "strong" UVB output; but when the intensity of the light is matched to that of sunlight, the risk of eye or skin damage will not be much greater than that of sunlight, either. Other lamps with almost identical spectra to this are the ZooMed Reptisun 5.0 tube and the Arcadia D3 Reptile 6% tube.
The ZooMed Reptisun 10.0 tube is similar but has a "stronger" output, with a greater proportion of its UVB spectrum in the shorter wavelengths, and a threshold wavelength of 293 - 295nm. This is more like tropical sunlight. As with tropical sunlight, there is a slightly higher risk of "sunburn" with a high dose (ie., extensive exposure or at close range) but we have never heard reports of eye or skin problems with a Reptisun 10.0 tube, or with an Arcadia D3+ 12% Reptile tube, which has a very similar spectrum.

The situation with the ExoTerra ReptiGlo lamps is slightly different. The ExoTerra ReptiGlo 10.0 has a different phosphor to the others - this is a classic "UVB" phosphor which emits much shorter-wavelength UVB and very little UVA. There is detectable short-wavelength UVB from about 287nm and a small but significant proportion of its output falls beneath the action spectrum for DNA damage. This lamp has been sold widely for several years and we have heard very few reports of eye or skin problems associated with its use; nevertheless, a small number of cases have been documented, usually (but not always) when a reptile has had close access to the lamp. The spectrum is not dissimilar in spectral power distribution to those lamps which did cause acute photo-kerato-conjunctivitis in reptiles several years ago, but it has a much higher threshold wavelength than those did, and has never to my knowledge caused problems on that scale. However, we can predict that this spectrum may carry a higher risk of skin and eye damage with extensive exposure, or at close range, than others shown on the chart.
The ExoTerra ReptiGlo 5.0 spectrum is different again. It has a slightly higher threshold wavelength than the 10.0 - around 290nm - and is midway between the ReptiGlo 10.0 and the Reptisun 10.0 in terms of the proportion that falls beneath the action spectrum for DNA damage. But it does have a high UVA output, which is unusual for a fluorescent reptile lamp. We have had a very small number of reports of reptiles and amphibians having problems which their keepers believed to be associated with use of the lamps, but considering the widespread sales of this product, the numbers are very small.

2. The intensity of the UVB (Lamp output).
Using broadband meters, we can estimate the intensity of the UV from each lamp.
The meter that we have now found to be most appropriate for measuring sunlight and reptile lamps is the Solarmeter 6.5 UV Index Meter (the version with a SiC sensor) http://www.solarmeter.com/model65.html
The sensor of this one responds very selectively to the shorter wavelengths, from 270 to 310-315 nm, which we know are most responsible for vit D synthesis and other photobiological effects on the skin.

The UV Index is a very useful scale to assess a lamp with, too, because the numbers are familiar to many people from "sun safety" reports, and give a useful sense of what is likely to be safe and "sunlike" and what is going to be frankly dangerous to eyes and skin of living things.
I now use the 6.5 UV Index meter, along with the spectrometer, to estimate a lamp's vitamin D3 potential - and safety.

Hand-held meters can never be "accurate" in the same way that a spectrometer can be. From a practical point of view, however, a Solarmeter 6.5 is more than adequate for assessing the value of a lamp and roughly matching its output with that of sunlight in a reptile's microhabitat in the wild.
It has proven effective in trials measuring vitamin D3 output in vitro, and new recommendations for suitable UVB gradients in vivaria are being developed based upon field studies recording the UV Index in the microhabitat of numerous species in the wild.

However, the most well-known broadband, hand-held meter that is affordable and yet reasonably accurate is still the Solarmeter 6.2 UVB meter
http://www.solarmeter.com/model62.html. This meter estimates the total UVB output (it responds to wavelengths from UVC right up to and including a little short-wavelength UVA at about 330nm) and gives a reading in microwatts per square cm (µW/cm²).
It's a good little meter, sensitive (down to 1µW/cm²) and reliable. I use my 6.2 all the time, to monitor decay and to work out the shape of a lamp's beam.

But you cannot compare solar readings with lamp readings, nor can you compare readings from different lamp brands, with a 6.2.
This is because as I've already mentioned, only the shorter wavelengths, up to about 315 nm, synthesize vitamin D3 in the skin; but the readings from the 6.2 meter include much longer wavelengths than this. In theory, if one lamp has a lot of its UVB in the shorter wavelengths, whereas another has a lot of its UVB in the longer wavelengths, they could both give identical readings with a 6.2, but the first lamp might be far better at producing vit D3 than the second one.

Nevertheless, many reptile keepers own these meters, so it's worth including readings from these meters, too.
And if we look at the ratio of UVI to total UVB (6.5 meter reading : 6:2 meter reading) we can learn something else very interesting. Since the 6.2 measures the whole UVB region, and the 6.5 measures only the shorter wavelengths, this ratio gives an indication of the proportion of short wavelength radiation in the spectrum.
This is very useful, because sunlight has far more long-wavelength UVB than short-wavelength UVB. The ratio is typically between about 1:50 and 1:65 (depending upon solar altitude and degree of cloud cover) ie., when the UVI reading is 1.0, the UVB reading is 50 - 65 µW/cm². Mid-day sun in the tropics may give ratios as low as 1 :35, since the higher the sun is in the sky, the greater is the proportion of short-wavelength UVB, and hence the lower the ratio.

With lamps, the ratio is almost always even lower, because most tend to produce a higher proportion of UVB around 300 - 315nm than sunlight. The lower the ratio, the more likely is the lamp to have a hazardous output. Brands of reptile lamp which give ratios of 1 :25 or higher have not, to my knowledge, ever been associated with any health problems in reptiles, however. Occasionally, lamps with ratios between 1 :12 and 1: 25 have been associated with problems, particularly if the lamp output is high and/or the reptile has close access to the lamp. This may be because cellular damage from UV is dose-related. On the other hand, all samples of fluorescent tubes, compact lamps and mercury vapour lamps acknowledged to cause eye and skin damage to reptiles, which I tested gave ratios between 1 :7 and 1 :12.
This sort of analysis is very useful if you don't have access to a spectrometer...

Here is a table with readings from the different meters, at various distances, from single samples of each brand of lamp. These are all new lamps, burned-in for about 100 hours so their output has settled to a reasonably constant level. The lamps I test are all 60cm, 18 - 20 watts. (Longer tubes will give slightly higher readings, not because their output at any point along their length is higher, but because the meter "sees" more of the lamp. This should not affect the chosen distance these lamps are placed above the reptile.)
Image

When choosing a lamp, you need to provide the appropriate level of UVB for the species you keep (based on its basking habits in the wild) whilst ideally, using a lamp that provides the closest match with natural sunlight, and the least risk to eyes and skin.
This probably means that for a sun-basking species, you will want to choose a lamp with a high UVB output, but the lowest ratio possible (ie., with as close a spectrum to natural sunlight as possible, in the UVB range.)

How much UVB at the basking site?
Of course, the next question you're going to ask is "but how much UVB do my turtles need?"
I don't have any measurements of the sunlight in which wild turtles bask, nor do I know anyone who has gone out there and measured it. We cannot possibly collect field data for every species. However, last year I co-authored a first attempt to develop a practical guide to this problem. Herpetologist Gary Ferguson and his team in the USA compiled a series of studies of the UV environments of 15 different species of reptiles (lizards and snakes) across the South and West USA and Jamaica, which were typical of a wide range of habitats. We published the paper in Zoo Biology. The reference is:
Ferguson, G.W., Brinker, A.M., Bucklin, S.E., Baines, F.M., Mackin, S.J. (2010). Voluntary exposure of some Western-hemisphere snake and lizard species to ultraviolet-B radiation in the field: how much ultraviolet-B should a lizard or snake receive in captivity? Zoo Biology 29(3) 317-334
http://www3.interscience.wiley.com/jour ... 6/abstract
In this, we developed what I call The Ferguson Zone concept. We believe that knowledge of the basking/ sun exposure habits of any species enables us to make a reasonable estimation of likely average UVB exposures. A suitable UV GRADIENT may then be provided in the captive animal’s environment, enabling the animal to selfregulate its exposure within the limits suggested by this assessment. It is vital that a GRADIENT is provided – there must be a full range of UVB levels from zero (full shade) to a suitable maximum based on an estimate of the animal’s natural microhabitat. This gradient must also be linked to light and radiant heat - after all, sunlight always provides all three together. In the wild, a reptile will move freely from places of full shade to areas having the level of exposure it desires, and it will move in and out of direct sunlight as it chooses. Once a suitable gradient is available, in captivity, the reptile will do the same.

All the UVI readings for the microhabitats at the time and place the reptiles were found were averaged. It was found that any reptile could be allocated to one of four "Zones" depending upon its basking habits. The four Zones are:
Zone 1: crepuscular or shade dwellers
Zone 2: partial sun or occasional baskers
Zone 3: open sun or partial sun baskers
Zone 4: open sun and mid-day baskers
The average exposure of Zone 1 species fell in the range between UVI 0 – 0.7, Zone 2 were in a range from 0.7 – 1.0.
Zone 3 animals had an average exposure between UVI 1.0 and 2.6, and Zone 4 were on average experiencing UVI 2.6 -3.5.
These averages were made over the entire activity period - they include higher readings when the reptiles were in full sun, and lower readings when in full shade. So each zone range might be seen as a suitable “mid-background” level of UVB for the species in question.

Different species of turtles will occupy different Zones, of course, and as yet we can only make "best guesses" as to what that Zone would be, for many species. I have seen red-eared slider turtles basking flat out in hot sunshine, sitting in long rows on the banks of outdoor zoo enclosures in Europe, however; so these are most certainly Zone 3. It is possible they might even be zone 4, but they survive well in the UK where there is rarely the opportunity for Zone 4 behaviour so if I was asked for my opinion, I'd suggest putting them in Zone 3 and aiming for a background UVI of between 1.0 and 2.6, across the entire basking area (with shelter and shade in other areas of the enclosure of course.)
A Solarmeter 6.5 will make life very easy for judging the correct placement of a UVB tube to achieve this. If a meter isn't a possibility, then the tables can provide a very rough guide.... but they have been made on single, small, new tubes with no reflectors fitted. A good aluminium strip reflector behind a tube can almost double the output beneath the tube. Two tubes together will double the output. Two tubes, fitted with reflectors which are carefully positioned, can almost quadruple the output.

For example: suppose I wanted to use the ZooMed Reptisun 10.0 tube. (These have proven reliable in all my tests; they appear to be high quality, made in Germany, and decay very little after the first two or three months, and so can be used for a full year before replacement is necessary.)
To get UV Index 2.0 from one, lone tube I would need to have it only 10cm above my turtle's back. Clearly that is unsuitable. It is much too close. If he reaches up, or climbs onto his mate's back he might touch the tube; and the glare in his eyes would likely be unacceptable.
But if I fit an aluminium reflector, I can double the output.... Using the chart, I can see that I will obtain the desired UVI 2.0 at somewhere between 20 - 25cm. (2 x 1.2 = 2.4) That would be an acceptable distance for most set-ups.
Or I could use a pair of tubes, set up together, parallel to each other, at about 20 - 25cm. (Best to keep them both directly above the basking zone. Nothing appreciates light, or UV light, shining sideways into its eyes.)
Or I could set up a pair of tubes with reflectors, at just over 30cm distance. (4x 0.7 = 2.8; but you won't get fully 4 times as much UVB.)

That's how it works in theory.
In practice..... we just do our best.

Frances