dramatically demonstrated by its firing 30 seconds after release. Launching shocks caused the baroswitch pen-arm to fall off its shelf, completing contact prematurely. In later flights, a time delay switch was placed in series with the baroswitch to prevent a recurrence of this action.
Flights 21, 22, 24, 26 and 27, although carrying altitude control devices, were flown to test gear for associated projects. Either no pressure reporting gear was carried or the data from modified gear proved unreliable. Hence few performance data charts are presented.
Flight 21, using a late-model General Mills 20 foot thin cell and an automatic ballast valve, is known to have lasted for ten hours, descending at Marietta, Oklahoma.
Flight 22, included an earlier model General Mills balloon with a high rate of gas leakage, and an automatic ballast valve. The ballast control kept the balloon aloft, but for only six hours.
Flight 24, including an automatic ballast valve, is believed to have maintained constant level, $1,000$ feet, for 122 minutes. It stayed aloft for at least $3\frac{1}{2}$ hours, when transmission ceased. The time-altitude curve is shown in figure 51.
Flight 27 employed a fixed rate of leak rather than an automatic ballast valve. The manual control did not provide sufficient ballast flow, accounting for the time-altitude curve shown in figure 52.
Flights 29 through 37 and flight 39 were undertaken to test the downwind launching procedure, to try for higher constant level altitudes, and to determine the feasibility of using the General Mills thin cells for frequent service flights. Flights 37 and 39 burst early. The former was released during a rainstorm and balloon failure occurred at the seams.
- 30 -