Thanks for your website again. And apologize for my poor English.

It seems that the ships did have an aispeed indicator (ASI) after all. In his book, Harold G Dick repeatedly gives very precise figures for the airspeed of the Graf Zeppelin and Hindenburg at various times. He mentions attempts to measure accurately the Hindenburg’s airspeed during flight test, using sensors mounted (suspended?) well clear of the hull and outside the boundary layer and says that they were suspected of mis-reading (see pp 106-107). One was a rotating-impeller type and the other a dynamic head type. In the glossary (p197) he mentions a “vibrating reed” type of speed meter. At no point, despite fairly detailed descriptions of the flight instruments, does he mention the actual control room or navigation room instruments themselves.
It seems that the vibrating reed indicator used a small windmill on a mounting out in the free airflow, similar to the windmill type electrical generators seen on some aeroplanes of the period and also, I believe, used on the Graf Zeppelin to provide power to the radio room. The faster the ship, the faster the windmill turned. A simple tachometer cable would have transmitted the rotation to the onboard indicator which worked by allowing a series of tuned steel “reeds” to vibrate in harmonic sympathy with the rotating mass. By noting which reed was vibrating, the rpm of the windmill impeller could be established and, since rpm were proportional to speed through the air, a more or less accurate reading of airspeed taken.
References
Dick, Harold G and Robinson, Douglas. The Golden Age of the great Passenger Airships Graf Zeppelin and Hindenburg, Smithsonian Institution, Washington DC, 1985. ISBN 1-56098-219-5
Seventh Annual Report of the [US] National Advisory Committee for Aeronautics, Including Technical Reports Nos 111-132. NACA, Washington DC, 1921. Available on Google Books

I have been doing some more research on DR navigation based on drift measurements. According to Harold G Dick, who was intimately acquainted with the Zeppelin company procedures, this was the normal method of long-distance navigation used and it was extremely accurate. Although the navigators had the means and training to use astro-navigation and an observation hatch for the purpose was provided in the nose of the ship, in practice astro was found to be unnecessary. Long-range radio navigation was unreliable and inaccurate in the 1930s. Hal Dick describes an 1800 mile oceanic leg in the Graf Zeppelin where the sole means of navigation was by hourly double-drift readings. At the end of some 25 hours’ run, landfall at Fernando Noronha island was precisely on course and within a few minutes of ETA.
The Zeiss drift sight in the LZ129’s navigation room appears to be an advanced piece of equipment for its day. I have found no detailed description of its use but a US Army Air Corps technical manual of 1940 describes what appears to be a very similar American equipment which is a near contemporary. It seems reasonable to assume that the capability of the two instruments and the techniques for their use were similar. In addition to accurate measurement of the drift angle (ie the angle between aircraft heading and track being made good over the surface) which results from the current wind at the aircraft’s altitude, the drift sight allows precise measurement of groundspeed.
Available descriptions of the ‘double-drift’ method all assume that the airspeed is known and that what is sought is the wind direction and speed, but nowhere can I find descriptive or photographic evidence of any form of airspeed indicator (ASI) on the Graf Zeppelin or Hindenburg. However, provided your drift sight allows accurate groundspeed measurement, it is still possible (though a little cumbersome) to use a modification of the method to find the wind direction and speed and thus your true airspeed, which can then all be used to determine the heading required to make good any desired track.
Firstly, it is essential to be able to measure accurately your height above the ocean. Both Graf Zeppelin and Hindenburg had Echolot sonic altimeters to do this but both also used a simple alternative using empty bottles of a particular brand of mineral water carried aboard. Tables had been drawn up of the ‘time of flight’ taken by a bottle to splash into the water below when dropped from the control car window. Provided the ship was over water, by day or night (with the searchlight) a stopwatch was used to measure a bottle’s time to impact and the table gave a precise height above the surface. Armed with this information, accurate calculation of groundspeed was possible by timing the passage of a surface feature through a known change in its angle of sight across the field of view of the drift sight and using simple trigonometry. In fact, tables would have been drawn up so that a very simple calculation revealed the ships groundspeed.
So here’s how the hourly procedure would probably have gone… Assume the ship to be mid-Atlantic, by night, heading SSE, say 202 degrees True, en-route to Pernambuco at an altitude of approximately 200m (650 ft). Firstly, the downward searchlight would be switched on and a bottle dropped and timed. The table is consulted and the barometric altimeters re-set to the exact altitude calculated. The navigator records the ship’s altitude and heading and peers through his drift sight to track streaks of foam on the surface. He records the angle between the ship’s heading and the path the foam follows across his field of view, the drift angle. He also records the time taken by some spot on the surface to pass between two cross-hairs in the drift sight eyepiece. In a few minutes, he will use these data to calculate his ‘track made good’ and groundspeed, but first, he calls for the ship to make a turn through approximately 45 degrees to the left. The rudderman spins the wheel and steadies the ship on a convenient heading, say 155 degrees True, while the elevatorman carefully maintains his altitude. The navigator records the new heading and, once the ship is stabilised, he takes another drift angle and groundspeed timing. Immediately, he calls for a turn 90 degrees to the right. The helmsman steers 245 True and the navigator takes another set of readings. As soon as he’s finished, he calls for the ship to return to her original heading and takes a final set of readings as a confirmation. The searchlight is switched off and the navigator sets to work with his tables and his drawing instruments. For each set of readings, he applies the drift angle left or right of heading, as appropriate, to calculate the track. For each, he enters the appropriate table with his height and timing to read off his groundspeed. He takes a blank sheet of paper and, from a point in the centre, which he labels ‘W’, he draws a line, ‘WA’ whose direction is the reciprocal of the first track measured in the drift sight and whose length represents the associated ground speed. From the same point he draws another line similarly representing his second (off to the left) and third (off to the right) tracks and groundspeeds, ‘WB’ and ‘WC’ respectively. From the end of each of these three, at ‘A’, ‘B’ and ‘C’, he draws a line whose direction equals the associated heading. If all has gone correctly, the three heading lines cross at a single point which the navigator labels ‘O’. He places a drawing compass point at ‘O’ and sets it to draw an arc through ‘A’. With luck and skill, the same arc passes exactly through ‘B’ and ‘C’. The radius of this arc represents his true airspeed and the length and direction of a line joining ‘O’ to ‘W’ gives him the speed and direction of the wind. Our navigator is now able to mark his navigation chart with the track and distance run since the last check to establish the ship’s current DR position. He is now able update the course to the desired landfall and to revise his ETA. Any new heading to steer is passed to the Officer of the Watch and the position, the ETA and a weather observation is passed to the radioman who will transmit a position report. Armed with the wind information and his true airspeed, the nav can calculate the correct heading to make good any desired track and predict the groundspeed along that track.
A footnote on the subject of airspeed: although there seems to be no evidence of an airspeed indicator on the ships, Harold Dick does quote a table allowing the crew to estimate static heaviness or lightness according to the pitch angle required to fly level. The table is drawn at a speed of “71.6 mph” (115 km/hr) which seems a very precise figure. Elsewhere, he quotes the Graf Zeppelin’s normal cruising speed as “72 mph” so it may just be that this is the speed achieved in smooth air at normal cruising engine rpm, rather than one indicated on an ASI.

References
Dick, Harold G with Robinson, Douglas H. The Golden Age of the Great Passenger Airships Graf Zeppelin and Hindenburg. Smithsonian Institution, Washington DC, 1985. ISBN 1-56098-219-5
(US) War Department Technical Manual: Air Navigation (TM 1-205). (US) War Department, Washington DC, 1940. Available on Google Books

I too noticed that no airspeed indicator is mentioned which surprises me.
Navigating by dead reckoning is a matter of solving a “vector triangle”. In order to navigate by dead reckoning, the navigator needs to know six things, the direction and magnitude of each of three vector quantities: wind direction and wind speed; the direction the aircraft is pointing (“heading” or “course”) and its speed through the air (“true airspeed”); the direction it is moving relative to the surface below (“track made good”) and its speed over the surface (“groundspeed”). Given four of these, he can find the other two, either graphically or by using trigonometry.
In the normal way of things, a leg of the journey starts from a known point and the navigator will have a wind forecast, showing the expected wind direction and speed (ie both the direction and length of one of his three vectors). He will know the direction he wants to go (his “desired track”) and usually also the airspeed his craft will maintain. What he still needs to discover is, firstly, the heading he must fly so that, despite the wind, his craft will follow the desired track and, secondly, his groundspeed along that track. Note that he does need to know what airspeed the aircraft will be making.
Now, provided the air is relatively smooth and an airship is relatively close to being in static equilibrium (“EQ” in airship jargon), a given power setting on the engines will always give a corresponding airspeed in straight flight. During flight testing of an individual ship, the airspeed corresponding to each power setting would be recorded. However, if the ship is noticeably “heavy”, it will have to fly along nose-up to provide some dynamic lift: conversely, if “light” it will fly nose-down to create a downward force. The resulting “angle of attack” causes drag because the ship is not pointing exactly into the oncoming local airflow. Theoretically, flight testing would allow a table to be drawn up of airspeed against power setting at the pitch angles which equated to various deviations from EQ. Zeppelin policy was to keep the deck angle within 5 degrees of horizontal whenever possible, so this effect would normally be small. Unfortunately, turbulence, updrafts, downdrafts and significant turns would also cause speed reduction: deflecting the huge rudders and elevators produces drag and slows the ship down both because of their own projection into the airflow and the fact that they also cause the ship to adopt an angle to the oncoming airflow. A large airship, with a mass of over a hundred tonnes, has considerable inertia and would take some time to recover from a single, short-term deviation of this sort. In bumpy conditions, with the rudderman and the elevatorman sawing away at their controls, the ship might spend long periods at some indeterminate, and probably varying, speed well below the “standard” figure for the power setting and deck angle.
It is clear from all of this that the navigator would find use for a direct means of measuring his speed through the air. The simplest form of airspeed indicator (or ASI) consists of a small metal plate at right angles to the airflow and mounted on a spring in such a way that the oncoming air pushes it back, tensioning the spring. The faster the airflow, the further back the plate is blown and the speed is read off from a scale which measures the spring’s deflection. You may just see such a primitive ASI attached to the inter-wing strut of a vintage biplane. More advanced ASIs use essentially the same principle in a slightly different way: they still rely on the pressure created by the speed of the air rushing past the aircraft to deflect a needle on a dial calibrated in units of speed, usually nautical miles per hour (knots).
There is, however, a further wrinkle. The Hindenburg had two different drift sights to measure the angle between her heading and the track she was actually making good over the surface. You can just adjust your heading to make good the desired track. Theoretically, since speed is simply distance divided by time, the drift sight can also be used to measure groundspeed by seeing how fast surface features pass through the field of vision when at a known height. Even over the featureless ocean, whitecaps on the waves could be used in this way since they are effectively stationary relative to the airship’s 70-80kt airspeed. Given the Hindenburg’s powerful searchlight, this technique could even be used at night. When I flew blimps in the 1980s we had no drift sight but could work out our groundspeed by seeing how long our shadow took to pass some spot on the ground below. Furthermore, an experienced observer can make a reasonable estimation of the surface wind speed and direction by examining clues such as the drift of smoke over land or the movement of the waves and streaks of foam on the surface of the sea. There would not be much difference between the wind at the surface and that at the passenger Zeppelins’ normal very low cruising altitude. And bingo! We have five of our six values – and the sixth does not really matter because the whole point of all this is to work out which direction you are actually going and how fast, your track made good and your groundspeed. Now you can signal Lakehurst with an ETA and tell Chef Maier whether he’ll need to provide an extra snack because you are running late.

That makes sense. I just saw this reply so a very belated thanks to you for your answer.

Here is my take, although I am only an armchair expert.

Hydrogen is cheap, but too dangerous. Helium is expensive, and a finite resource! When helium is released, it floats to the top of the atmosphere, and on warm days the average molecular velocity exceeds escape velocity, and the helium “evaporates” off of the earth and out into the cosmos. It is too precious a resource to waste on airships.

The biggest vulnerability of airships seems (to me) to be their inability to withstand the stress of turbulence in storms. Blimps never experienced this limitations. I propose that the entire aluminum structure of the airship be replaced with air-inflated rings and longerons, a structure like a swimming pool toy, able to take the roughest treatment. And make it out of many independent cells, so a single puncture would not threaten the whole structure.

Could be done, but the limiting resource is helium. If we could handle hydrogen safely (for peace time, not military applications) it might be feasible.

Just MY take. What do I know?

Counterbalance and counterweights — the axis of rotation for the planes was not their leading edge, so some airflow was working to hold all the fins centered, and counterweights in the elevator planes would further assist. Counterbalance is easy to see on most airplane control surface and even on boat/submarine surfaces, so we can think of that as a very old trick indeed.

There was also a question regarding the British being able to shoot down Zeppelins during WW1. A special exploding machine gun round was developed that contained nitroglycerin; the round was designed to tear a relatively large hole in the gas bag, allowing enough hydrogen to escape to create a bubble of combined hydrogen and air. Machine gun magazines were loaded, alternating the explosive with magnesium tracer rounds. The explosive rounds created holes that allowed hydrogen and air to mix. The tracer rounds followed essentially the same trajectory as the explosive rounds and ignited the hydrogen/air mixture on their way through. The machine gun rounds had to be concentrated onto one point, and it took several rounds to ignite the hydrogen/air mixture.

I couldn’t figure out how the nitroglycerin explosive rounds worked; not exploding when shot from the gun, and exploding at a point where the resulting hole would allow hydrogen and air to mix.

— Tim