house. The temperature changes in the two limbs, on their being moved in and out of the greenhouse, were noted for several days. The behavior was practically the same as when the limbs were attached to the tree. The dead limb was then soaked in water for some time and the temperature changes of the two limbs again observed during both rising and falling temperatures. These observations show that the rise and fall of temperature of a dead limb is very appreciably checked on approaching the freezing point. In short, a soaked dead limb behaves like a live one, as far as temperature changes are concerned. There seemed also to be some indication that the temperature of the soaked dead limb, after having once passed below the freezing point, falls faster than that of the live limb below the same point. The later comparisons of the temperature changes of a live limb and a dead one not soaked are illustrated by the diagram for January 26, 1897. (Fig. 3.) A like comparison of a live limb and a soaked dead one is given by the diagram for February 26, 1897. (Fig 4.) In both these cases the limbs were placed out doors at 9 a. M., after having first acquired a uniform temperature in the greenhouse. EXPLANATION OF PLATE I. (FIG. 1.) Sept. 9, 1896. Curve No. 1 shows the changes in the temperature of a thermometer in air. No. 3. Same at center of dead limb. No. 4.-Temperature near surface of live limb. No. 5.-Same at center of live limb. Limbs shaded till 1:15 P. M., in sunshine till 2 P. M. Sky cloudy remainder of day. (FIG. 2.) Dec. 12, 1896. No. 1 shows the temperature of air. No. 2 shows the temperature of surface of dead limb. (FIG. 3.) Jan. 26, 1897. No. 1 shows the temperature of air. No. 2 shows the temperature of surface of dead limb. No. 4 shows the temperature of surface of live limb. (FIG. 4.) Feb. 26, 1897. No. 1 shows the temperature of air. No. 2 shows the temperature of surface of soaked dead limb. No. 4 shows the temperature of surface of live limb. DATA AS TO WIND-DISTRIBUTION OF SEEDS. EDWARD M. HUSSONG. Few matters are of greater economic importance than that of how and to what extent the seeds of flowering plants are distributed by winds. To the ecologist and phyto-geographer it is one of no moderate interest, as geographical distribution, though local, is in each of these lines of modern research no inconspicuous part; to the student of local floræ it accounts partially for the unexpected appearance of species whose natural habitat has been assigned remote from his district of special work; while to the agriculturist it affects vitally his success or failure on the farm. No available data touching directly upon the subject have been tabulated; perhaps none have been collected. Under my direction the students of the Franklin High School have carried forward for the past three years a series of field collections and laboratory cultures that have opened the problem at least to wider investigation. Our work briefly is thus: Lee We have made exposures of collecting traps in the various situations afforded us on the high school campus, around the homes of the students, and in the open prairies and fields. ward and windward positions were selected when previous indications foretold the direction of the wind; the value of these different positions is apparent from the widely different results obtained thus from the same period of observation. The different situations,. as campus, barnyard, prairie, ravine, field, etc., give lawful variations which the observer can easily account for. The traps used were deep tin cans, anchored fast by means of stout stakes to which the cans were securely wired in a manner allowing of easy detachment to remove the contents. Winds were grossly classified into four groups: breezes, local winds, high winds, and continued gales; such grouping being based upon the force and continuance of atmospheric movement. The trash and dust were carefully removed from the cans and labeled as to location, date, length and strength of wind. A specimen label is as follows: "Harry Davis; open pasture south of Franklin, one-half mile from tilled soil; continued gale, two days; Oct. 10-11, 1894." In many instances dates were not recorded, but the general season-period is known. These omissions have so affected the calculations that no definite dates can be given in the tabulation; the early or late season-period, however, is fairly accurate. Vegetable mould, calorized to prevent extraneous growths, was added to the collections and under fair conditions the seeds contained therein were germinated. Many of the plants grew to maturity, others developed sufficiently to reveal their identity, a few were classed "unknown." The tabulation of results is not as satisfactory as could be wished, but it suffices to show the general trend of the investigation. The following table shows calculations obtained from fortyeight collections, 1893-4-5-6, taken during |