88 



NATURE 



[September 23, 191 5 



magnetometer method might be used more than it is 

 at present under suitable conditions ; I wished to 

 direct attention to its practical limitations. 



The Reviewer. 



Nodules on the Intermediate Bladderwort. 



I RECENTLY had an opportunity of examining living 

 specimens of a very rare British plant, the inter- 

 mediate bladderwort {Utricularia intermedia), and 

 found that the leaf-bearing stems ended in a solitary 

 terminal knob or nodule, the largest being the size 

 of a small peppercorn. What are these Icnobs? 

 Are they autumn plant-buds that will ultimately grow 

 into plants? This seems to me a possible solution; 

 but (my edition of) Bentham does not mention that 

 the plant propagates itself in this way, nor are the 

 knobs figured in the companion volume of illustrations 

 by Fitch and Smith. Harold Evans. 



Llanishen, Cardiff, September 17. 



NOTES ON STELLAR CLASSIFICATION. 

 III. 



TN the year 1890 a photographic study of the 

 -*- spectra of stars vi^as commenced by me at the 

 Solar Physics Observatory at Kensington. The 

 object of the investigation was not so much to 

 make a spectroscopic survey of the stars generally 

 as to examine with considerable dispersion the 

 spectra of the brighter individual stars. Up to 

 the end of the year 1900 there had been accumu- 

 lated a large number of spectra, a catalogue of 

 which was published in the year 1902. This con- 

 tained 470 of the brighter stars. 



In this catalogue the stars were classified ac- 

 cording to their spectra after a minute inquiry of 

 the lines due to each of the chemical "elements " 

 involved. 



The spectra of 105 of these 470 stars were 

 photographed by Dr. F. McClean at the Cape 

 Observatory. 



Details of this classification were given in the 

 publication above mentioned, but a brief summary 

 may here be given. The stars were arranged in 

 two main groups, one in which the stars were 

 getting hotter ("hotting"), and the other in which 

 they were cooling. The differentiation of these 

 two groups was based, not only on their chem- 

 istry, but on other differences observed on the two 

 arms of a "temperature curve." 



On the temperature curve the stars increas- 

 ing their temperature were located on the left- 

 hand branch or ascending arm of the curve, 

 and the cooling stars on the right or descending 

 arm. The stars of the highest temperature were 

 located at the apex of the curve. 



The two main divisions of stars, i.e. those 

 getting hotter and those getting colder, were each 

 classified according to their chemistry and placed 

 at their respective levels on each side of the tem- 

 perature curve. At equal levels on each arm the 

 stars were considered to be of the same tempera- 

 ture, each group being designated by a name 

 derived generally from that of the " type star " 

 which had been selected to represent that group. 



Further experience with this method of stellar 

 NO. 2395, VOL. 96] 



classification led to the detection of criteria which 

 have suggested the interpolation of additional 

 groups, and those have now been embodied in the 

 general scheme as given below representing the 

 classification in use at present, and to it has been 

 added the Harvard scheme of classification of the 

 type stars employed by me. 



The main routine work at the Hill Observatory 

 at Sidmouth, since the McClean telescope was 

 brought into use in September, 1913, has been a 

 spectroscopic survey of all stars down to the fifth 

 magnitude other than those published in the 

 above-mentioned catalogue ; incidentally the 

 spectra of a number of stars fainter than magni- 

 tude five have also been photographed. 



This programme of work is especially fitting for 

 this observatory, because the same instrument is 

 in use which Dr. Frank McClean employed in his 

 spectroscopic survey of stars equal to or brighter 

 than magnitude 3*5, the telescope having been 

 presented to the observatory by his son, Frank 

 Kennedy McClean. 



The McClean instrument consists of an equa- 

 torially mounted twin-telescope with apertures 

 of 12 and 10 inches, and focal lengths of 134 and 

 150 inches respectively. The former is fitted with 

 an objective prism of 12 inches aperture and 

 20° angle, and is mounted in a metal frame in 

 such a way that it may be moved clear of the 

 objective. In the focal plane is placed a camera 

 fitted with a swing back by Messrs. Hilger. This 

 carries a plate-holder for plates 6| x 4! , but an 

 adapter has been inserted to carry plates 4J x 3J 

 which are in use. The dispersion on the photo- 

 graphic plate between K and H)8 is equal to 

 28 mm. or 927 Angstrom units. 



The telescope, built by Sir Howard Grubb, of 

 Dublin, is driven by means of falling weights, and 

 is fitted with electric control movements regulated 

 by an electric pendulum. The governor is of the 

 heavy ring pattern, and is adjusted by means of 

 a cam. The declination circle is electrically 

 illuminated, and viewed with the aid of one of 

 two reading telescopes from the camera end. In 

 addition to two finders of 4-inch and 2-inch aper- 

 ture fitted to the lo-inch tube, a 3-inch finder 

 deviated to the angle of the prism has been 

 specially attached to the tube of the 12-inch; this 

 serves the purpose of observing the amount o( 



