This Chinese company "thinks" that they have created a self ballasted metal halide lamp!

http://townlite.en.alibaba.com/productshowimg/321700244-213165925/400W_self_ballast_metal_halide_lamps.html

    Either it is a Mercury arc tube or its efficacy is so low that one need not bother about it.

It will not last long or blow up on the first start. It looks poorly made.

Even when that is not as good quality, I would be interested to have some, simply as an example.
But it does not have to be that bad, giving the specifics the combination of an incandescent and MH in a close proximity.
The arctube have to be somewhat larger, as it is heated from the very close placed ballast lamp, while it should have the same cold-spot temperature as the stand-alone MH.

These looks like halogen ballasted mercury lamps.
But there are two points of thinking:
1. These pictures may display different products then that the company want to sell (Common phenomenon in online lighting shopping site, especially american and chinese sites).
2. This may be a fraud site.

Colin has one of these and has posted a couple of pictures...

http://www.lighting-gallery.net/gallery/displayimage.php?album=229&pos=16&pid=72276
He has not included the spectrum though, so It could simply be a self ballasted mercury, the arc tube looks like a mercury arc tube.

I think you can't get top performance MH on a resistive ballast - too much restrictions on the design : no restrike after zero crossing of the AC, need low starting voltage....

But it can be made along the lines of SBMV just with a bit of halide added

But the arc tube is clear.... where are the halides ?

To make a MH lamp ignited directly from the mains 230V, it have to be a probestart MH lamp with one or two auxiliary electrodes and a neon-argon penning inside the arctube.

I would guess this concept would give way superior performance over the traditional mixed lamp. The main reason is the decoupling of the outer atmosphere from the incandescent atmosphere: With the incandescent having it's own bulb, the inner pressure and composition could be tailored to fit the incandescent requirements (so even high pressure halogen cycle,...), while the outer pressure could be kept low and with pure nitrogen, so the arctube does not lose as much heat. Moreover the heat from the halogen is clearly utilized to aid the arctube to retain it's temperature...

It would be good to measure the spectrum, to identify what is inside the arctube.
I had a look to the (most likely the real) manufacturer's info (on Alibaba.com) and the specified total lumen output (12000lm) is too high to be from only mercury discharge lamp, even when both the arctube and the incandescent would be perfectly optimized as per individual devices:
Assume 60% power goes on the incandescent part (so 150W, assuming best case 24lm/W mean 3600lm from the halogen)
So for the arctube remain something about 100W.
With mercury arctube would be really optimistic to achieve ~40lm/W, what would mean 4000lm.
So with the mercury arctube, the best case total output would be about 7500lm.
To reach the specified 12000lm, the arctube would have to emit about 8400lm, what would mean 84lm/W, something really needing some other active components than mercury, while the others are known to be usable only in the form of halides.

The halide composition does not have to radiate in red and yellow, as the lamp is specified as 5500K, while the ballasting incandescent lamp provide quite enough red and yellow output. So sodium won't be necessary at all. That mean, the most prominent corrosive agent for the quartz being gone, there is no need for the starting probe bypass switch, normally found in nearly all probe start MH's (hence the simple resistor only)
What would be needed are some strong green and blue radiators (Thallium, Indium, Hg?), while all of these colors mean a lot of lumens per relatively low real radiated power, so quite easy to reach the rather high efficacy...