Popular Electronics
The following comments relate to an article, How to make a Lava Lamp, by Ralph Hubscher, from Popular Electronics magazine, March 1991, p. 31 (4). Gernsback Publications.
   You are encouraged to find the original article and work from that. These are notes of a speculator.

Several non-water-soluble chemicals fall under the category of being "just a little bit heavier" than water, and are still viscous enough to form bubbles, not be terribly poisonous, and have a great enough coefficient of expansion. Among them: Benzyl alcohol (Specific Gravity 1.043 g/cm3), Cinnamyl Alcohol (SG 1.04), Diethyl phthalate (SG 1.121) and Ethyl Salicylate (SG 1.13). [The specific gravity of distilled water is 1.000.] Hubscher recommends using Benzyl Alcohol, which is used in the manufacture of perfume and (in one of its forms) as a food additive. It can be obtained from chemical or lab suppliers such as VWR Scientific Products. An oil-soluble dye is nice to color the "lava"; Hubscher removed the inards of a red marker and soaked it in benzyl. Apparently it worked great. [Benzyl alcohol is "relatively harmless", but don't drink it, and avoid touching & breathing it.] Hubscher found that the benzyl and the water alone didn't do much, so he raised the specific gravity of the water slightly by adding table salt. A 4.8% salt solution (put 48 grams of salt in a container and fill it up to one liter with water) has a specific gravity of about 1.032, closer to benzyl's 1.043. We all know that salt will cloud the water and that's just not cool. So you might want to experiment with other additives. (Antifreeze?, Vinegar?)

Some More Theorizing
Submitted to a newsgroup in '95 by Bill Beaty

"The "lava" smells like gasoline. It seems unsafe, but could it be a mixture of paraffin wax thinned by gasoline? By itself this wouldn't sink in water, but by varying the amounts it might be possible to adjust the viscosity and melting point. The mixture might not be very flammable if only a small amount of gasoline is needed. Use extreme care if you experiment with flammable materials!
    Another clue: after a few years a LavaLamp fails because the 'lava' separates into two masses, one which floats and one which sinks. This suggests that one ingredient may be a finely powdered solid. Over time this solid settles to the bottom of the 'lava', and eventually part of the 'lava' becomes too light to sink. The powder would need to be very fine, and should be dense enough to sink the lava, but not so dense that it settles within the 'lava' quickly. Could it be something as simple as talc? Or maybe it's very fine silica."

If you have any ideas or questions, just post them in the Science Lab discussion of the Lava Line.