[Illustration: Fig. 8--"Hon-woo-et-low" or "Copper Jars Dropping Water"--Canton, China] The most interesting of all water clocks is undoubtedly the "copper jars dropping water," in Canton, China, where I saw it in 1897. Referring to the simple line sketch, which I make from memory, Fig. 8, and reading four Chinese characters downwards the translation is "Canton City." To the left and still downwards,--"Hon-woo-et-low," which is,--"Copper jars dropping water." Educated Chinamen inform me that it is over 3,000 years old and had a weather vane. As they speak of it as "the clock of the street arch" this would look quite probable; since the little open building, or tower in which it stands is higher than surrounding buildings. It is, therefore, reasonably safe to state that the Chinese had a _weather and time station_ over 1,000 years before our era. It consists of four copper jars partially built in masonry forming a stair-like structure. Commencing at the top jar each one drops into the next downward till the water reaches the solid bottom jar. In this lowest one a float, "the bamboo stick," is placed and indicates the height of the water and thus in a rude way gives the time. It is said to be set morning and evening by dipping the water from jar 4 to jar 1, so it runs 12 hours of our time. What are the uses of jars 2 and 3, since the water simply enters them and drips out again? No information could be obtained, but I venture an explanation and hope the reader can do better, as we are all of a family and there is no jealousy. When the top jar is filled for a 12-hour run it would drip out too fast during the first six hours and too slow during the second six hours, on account of the varying "head" of water. Now, the spigot of jar 2 could be set so that it would gain water during the first six hours, and lose during the second six hours and thus equalize a little by splitting the error of jar 1 in two parts. Similarly, these two errors of jar 2 could be again split by jar 3 making four small variations in lowest jar, instead of one large error in the flow of jar 1. This could be extended to a greater number of jars, another jar making eight smaller errors, etc., etc. But I am inclined to credit our ancient Chinese inventor with the sound reasoning that a human attendant, being very fallible and limited in his capacity, would have all he could properly do to adjust four jars, and that his record would average better than it would with a greater number. Remember, this man lived thousands of years before the modern mathematician who constructed a bell-shaped vessel with a small hole in the bottom, and proportioned the varying diameter in such a manner that in emptying itself the surface of the water sank equal distances in equal times. The sand glass, Fig. 9, poetically called the "hour glass," belongs to the water-clock class and the sand flows from one bulb into the other, but it gives no subdivisions of its period, so