But why? Why must it be that way? Impy delighted my black and shriveled little heart by challenging my assumption on twitter (seriously, I was thrilled to see her contradicting me.) So I thought that those of us with creative bents, too much time on our hands, and a complete lack of fear when it comes to mixing fantasy and physics might have a little fun exploring. So here's the deal.
Design an airship for a pirate or naval officer and rationalize it. It can be in one of three catagories: Real World (meaning it has to be doable within the laws of physics as we know them); Steampunk (meaning you have to use a lifting gas of some sort but you can use some sort of applied phlebotenum to lighten materials, exaggerate the lifting power of a lifting gas, or harden materials); and High Fantasy (meaning the use of magics and/or fantasy creatures to get the job done.) Then we all can get our panties in a bunch arguing about how the idea would never work, following which we don top hats, monocles, and go get a pint at the closest public house.
Ready? Go!

(A few useful bits of information:
Volume of a sphere: V=(4/3)πr^3 Surface area of a sphere: A=4πr^2
Volume of a cylinder: V=πr^2h Surface area of a cylinder: (end piece) A=πr^2 (side) A=2πrh
Volume of an ellipsoid : V=(4/3)πabc Surface area of an ellipsoid (approximate): 4π((a^pb^p+a^pc^p+b^pc^p)/3)^(1/p)
Where r=radius; a, b, c; equal each cardinal radii on an ellipsoid; and p=1.6075.
For comparison purposes, the Hindenburg had a volume of 7,062,000 cu. ft. and used hydrogen to achieve a lift of 510,000 lb and weighed 470,000 lb without gas. The U.S.S. Acron had a volume of 6,500,000 cu ft and used helium to achieve a lift of 400,000 lb.
This gives us a basic 16 lbs per 1000 cu ft. using hydrogen and 13 lbs per 1000 cu. ft. using helium.
Feel free to correct me on these.)