I've observed for some months now the science sites stats on Stack Exchange after the closure of theoreticalphysics.se. I was glad that after the successful launch of biology.se, chemistry.se also finally went into public beta as the connecting scientific branch between those two. Its Area 51 stats, especially the committer roles with 40.2% academics were similar to biology and physics and much better than for example Economics or Cognitive Sciences.
Nonetheless, some of the mentioned sites have trouble in gaining traffic/questions, economics and theophysics were already closed, scientists don't have much time and are badly paid, they quickly begin to ask how much time should I invest in this beta experiment (a side note, naming these sites officially beta sites is kind of counterproductive for professional users. Luckily the ugly old beta school block CSS layout was removed too; personally I don't care as I use the Stylish Firefox addon, but the majority does).
To me it comes down to critical mass gained during the first 2-3 months of public beta and inflow of new user from WITHIN the Stack Exchange network for those science sites, because:
the growth via users from the outside googling and linked to Stack Exchange does not work for tricky scientifc questions (10 researchers will formulate the same question on molecular-genetics in 10 different manners compared to a specific code line or software install tutorial on Stack Overflow or Super User).
even the Stack Exchange registered user on science sites don't seem to notice that a new science site has started they were not comitted to (this is proven by the fact that the top rep users of physics, chemistry, biologly and computationalscience.se very seldomly participate or register at all on those other sites, despite the huge thematic and research overlap). I thought chemistry.se would have it much easier than a site like cognitivesciences.se (with very few psychology or neuroscience students on a programmer site) to attract new users and questions as connecting branch from biology.se and physics.se. But, most seem to not have registered at all when browsing the user pages. Many of the committers probably forgot the site after two years of commitment process (of course, understandable, probably a proposal should be set to a max. time (6 months) from proposal to private beta to guarantee critical mass).
I also observed Quora and subscribed to the chemistry and biology feed, 200 and 250 entries in one month (similar as expected, both huge fields). Quora has a different system (categorizing and following by tags and therefore will never reach the high-quality of dense expert communities grown via Area 51 like mathematica.se or physics.se), but every scientific topic is basically existing, you can ask questions and it probably gets answered.
Chemistry.se fullfilled all Area 51 requirements. I don't blame them; my hypothesis is that the Area 51 growth model doesn't work for such sites anymore. This and similar science sites are likely to gain no further growth as they never reached the critical mass during the first beta months like biology.se or physics.se (I don't think lowering the requirements is the way to go as a healthy community needs min. 3-5 new questions/entries a day to keep user visiting that place). Cognitivesciences.se is now 6 months in beta, linguistics.se even longer, the views/day constant around 200, no one seems to be able to google those sites or stumble over it inside the Stack Exchange network, despite of seeding the sites with quality content for some weeks now.
What can be done here? Looking at Quora, I think there are some lessons to be learned:
Many new users come to Stack Exchange via the hot questions front page (Quora has a similar followed tags user front page). Unfortunately the algorithm determining the hot questions is in my view pretty unfair. Posting highly voted Q&A from cognitivesciences, biology, chemistry.se from time to time is IMHO a must. Questions in these branches often yield only one good correct answer compared to the discussion threads on Science Fiction and Fantasy's Star Wars or Star Trek 4th or 5th redundant question. The more answers the higher the likelihood to be hot. Quality over quantity, is this not the philosophy here? The current algorithms only favors sites in advertising that already are graduated and have no growth problems and where users often browse the specific front page anyway or are subscribed via RSS. But a natural scientist visiting Stack Exchange for the first time, how are the current hot questions interesting to him?
offering feeds to science hot questions (containing hot questions in all scientifc Stack Exchange sites. Currently you have to click this together somewhere, browse all science sites, search the RSS link, import all feeds in your reader, etc. This has to be easier).
science, culture, programming filtering/categorizing buttons on the Stack Exchange hot questions front page, so you can easily browse the hot questions as a new unregistered user there interested mainly in one branch of Stack Exchange sites. Currently English, math, sci-fi, and programming is the standard hot stuff; how do you want to attract non-programmers to Stack Exchange this way?!
maybe showing chemistry.se questions tagged "material science" also on physics.se or at least in related questions sidebar, when the same tag exists on physics.se. Many questions on Quora are tagged physics/chemistry or biology/chemistry as the overlap is huge!
That are brainstormed ideas, there are probably better. But I think something has to happen, either a drastic change of the Area 51 system for non-pogrammer topics or advertising of such less-frequented new science sites not related to programming.
Otherwise, scientists you try to attract to fill such ghost sites with content will ask why should I further or at all participate in that experiment. Personally, as a physicist, if there is no way to build and maintain a chemistry.se site on Stack Exchange as connecting branch, then the current system and/or philosophy of Stack Exchange has a huge bug. But maybe I have a misconception and Stack Exchange is designed and created as a community around programming and programmers' hobbies and topics that fail to connect to such are doomed or have no future here?