Better Fruit, Nov. 1911, p. 42. A torch can be made from a stick having one end wrapped with cloth or asbestos soaked in kerosene. TANKS Where orchards are situated at easy hauling distance from the railroad, large wagon tanks of a capacity sufficient for at least two heatings will be sufficient without the construction of a storage tank upon the ranch. This minimum capacity is advisable for the reason that frosts usually occur by twos, the frost of the second night sometimes being more severe than that of the first, and hauling at a critical time such as this, when heaters must be refilled in good season for the second night's work, would be inconvenient and possibly disastrous. At ranches far removed from the railroad or where greater convenience in handling the oil is desired, a storage tank is necessary. But in either case, whether wagon tank or storage tank be used, a small faucet should be placed at the lowest point of the tank to insure the complete drawing off of the surplus water, which causes explosions and loss of oil when allowed to accumulate in the heaters. The size of tanks can be easily computed on the basis of the number and capacity of the heaters employed. A large but safe estimate would be three gallons per heater for each possible heating, or six gallons per heater in readiness for each series of frosts. Unless the orchard is well sheltered by windbreaks and the degree of frost is slight, one heater or even more should be provided for each tree. THE FORECASTING OF FROST One of the chief deterrents from orchard heating is the feeling that frost occurs without giving sufficient warning of its approach, and the consequent necessity, where the Weather Bureau does not distribute frost forecast bulletins, of arising frequently during the night, or maintaining a watchman throughout the season of possible frost cccurrence. In the case of the small orchardist who does his own work, this is exhausting; to the owner of the large orchard, who employs labor, it is expensive, to say the least. A careful study of the conditions under which frost occurs and of the general changes of weather should enable an alert observer to determine with increasing accuracy the nights on which frost almost certainly cannot occur and those in which its occurrence is certain or highly probable. The following brief statement of frost phenomena will lay the foundation for detailed observation: Changes of temperature accompany changes in barometric pressure. These changes are more or less gradual and occur in cycles of several days duration, but have no definite length. On mountain tops the temperature and pressure rise and fall together; that is, when a storm occurs, the temperature falls, and, when fair weather returns, it rises again. The surface conditions in the valleys apparently are the reverse of those on mountain tops. When the barometer falls the temperature rises, and when it rises the temperature falls again. More specifically, however, the temperature rises immediately preceding a storm and falls as the storm. progresses, reaching the minimum during the period of clear calm. weather immediately succeeding the storm. This low temperature is probably due in part to the influence of masses of cold air whose presence on mountain tops in times of storm has already been noticed. Such low temperatures usually do not long persist, but soon yield to more moderate temperatures. This general fluctuation in temperature is quite distinct from the diurnal fluctuation, which usually attains its highest or maximum in the early afternoon and its lowest or minimum temperature about sunrise. The diurnal fluctuation is due to the heating of the air by the sun's rays, and the subsequent cooling after sunset. This cooling or inversion of the temperature is due to radiation, which is greatest at the ground and gradually diminishes in intensity above the earth's surface. Hence, the lowest temperatures caused by radiation are found at the surface of the ground, and so rapidly does the intensity of radiation diminish that the temperature at the height of the lower branches of trees may be several degrees higher. Also the lowest temperatures due to radiation are found in the deepest parts of the valleys, while the corresponding temperatures on the elevated slopes of the valleys are higher. The amount of the diurnal fluctuation (or range) in temperature will depend upon the velocity of the wind and the amount of cloudiness at the time. The presence of wind or clouds during the day will materially lower the maximum temperature that would otherwise occur, while the presence of these same elements during the night will raise the minimum temperature above normal. This is due to the fact that the clouds that shut out the sun's heat by day likewise shut in the earth's heat by night, while the wind that mixes the warmer air with the colder by day likewise mixes the colder with the warmer by night. In forecasting frost these two fluctuations in the temperature should as far as possible be kept distinct from each other. They can, of course, be determined best by means of a thermograph or recording thermometer, or by a pair of special thermometers-a maximum and a minimum-which indicate respectively the highest and lowest degrees of temperature that have occurred since they were last set. A common thermometer that indicates only the temperature prevailing at the moment of observation, unless read at frequent intervals, is quite inadequate for this purpose. An analysis of two thermograph records made during the actual occurrence of frost will make these points clear. Plate IX contains a copy of the temperature trace made at Plum's Orchard during the week beginning April 24, 1911. It will be noticed that the maximum temperature decreased each day until Thursday, when it rose again with utmost regularity until the end of the week. This variation is part of the cyclonic fluctuation first referred to, which accompanies storms. The minimum temperature, like the maximum, also fell and rose again, but not with the same unbroken regularity that attended the fluctuation of the maximum. The irregularities in the fall and rise of the temperature during the week were due to wind and cloudiness, which lowered the maximum and raised the minimum beyond the normal. The dotted line represents the cyclonic fluctuation that probably would have occurred had the sky been cloudless and the air moderately calm. However, such ideal fluctuations do not usually occur, for the cyclonic changes in temperature are due to disturbance in the equilibrium of the atmosphere, and the characteristic phenomena accompanying atmospheric disturbances or storms are winds and cloudiness. Extending either side of this dotted line appear the diurnal fluctuations of the temperature, due, as has been said, to the heating of the air by the sun in the day time and to the radiation of heat at night. This diurnal fluctuation, where unhindered, is approximately uniform for periods of considerable length, and can be determined within practicable limits by comparing the amount of fluctuation for several days when the weather is fair. The record already studied will serve this purpose sufficiently well. The amount of fluctuation during the week, computing from the maximum of one afternoon and the minimum of the following morning, is as follows: Hence the fluctuation sometimes amounts to as much as 34 degrees and sometimes to as little as 20 degrees. When the weather |