To any one who has had the care of a hothouse during one of our terrible northern winters, the mere mention of a frost suggests hours of anxious watching and apprehension. I have sometimes thought that the pleasure taken in my hothouse during its first winter was more than counterbalanced by the constant anxiety and fear lest by some unforeseen circumstance the little silver column should drop below the fatal point, 32°, and in an hour the result of the patient labors of months, and the objects of my 'especial pride, should turn to blackness and decay. This fear grew to become a perfect night-mare, and my slumbers were frequently disturbed with visions of plants frozen and covered with ice. One cold night, after having made three visits to the hothouse to see that all was right, I resolved that something must be done, and commenced soon after to investigate the subject of electricity and its adaptation to burglar and fire-alarms. After a series of futile experiments, I obtained a small fiat rod or bar about twelve inches long, three-eighths inch wide, and one-sixteenth inch thick, formed by a thin piece of brass and a similar piece of steel fastened securely together.

This was suspended by one end being firmly fastened in a small block of wood placed on a board, and the rod was so placed that the free end could swing back and forth and just clear the board. On both sides of the bar, and about one inch from it, near the free end a thumb-screw was placed, so that if the bar be moved it would strike the ends of the screws. The end of the bar was fastened, and the screws were so arranged as to be easily connected with wires. Now the well-known law of physics that "heat expands and cold contracts" is true in metals, but in a different degree, and by consulting the tables of the expansive qualities of metals it will be found that steel and brass are widely different in this respect, and in the arrangement above described it was found that when heat was applied the brass expanded more than the steel, causing the rod to bow, and the free end to swing in the direction of the screw on the. steel side of the bar; and the application of cold caused the brass to contract more than the steel, and the bar to swing in the other direction. By testing this machine in various different temperatures, it was easy to make a scale, and to place the thumb-screws so that the end of the rod would touch them at any given point of temperature.

Then obtaining a common electric call bell, and a battery, such as the telephone companies use (any good battery will answer, but this one is always in order), the bell was placed in my chamber, and the battery and machine previously described placed in the coldest part of the hothouse. One pole of the battery was connected by a copper wire with the bell, and from the bell the wire was carried out to the machine and connected with the end of the bar that was fastened, and the thumb-screws connected to the other pole of the battery. I then placed the screw on the brass side so that it would come in contact with the bar in case the thermometer should reach 40°, having previously found that the slightest contact would complete the circuit and ring the bell. After waiting about two weeks without hearing anything from the apparatus, I was startled from my chamber by the ringing of the bell, and hastened out to find that a sudden and severe change had lowered the temperature to 38° in the hothouse, and but for the increased fire that was added the plants would have suffered a bad chill, if not frozen before morning. This has been kept in operation for two years, and has several times saved my plants from total destruction, or at least from great injury.

By adjusting the screw on the steel side of the bar too great heat is easily detected. Since thoroughly testing the alarm have come to put great confidence in it, as it can be regulated to within a single degree, and while absent or asleep my anxiety is reduced to a minimum.