IRIS. 



113 



tmijarrnfu the very high reactions with 



pyrogalln in ill, nitru- acid, sulphuric ami, h\dr. hluru- 

 potassium hydroxide, potassium iodide, potassium 

 sulpha yanatc. wdhUB hydroxide, sodium sulphide, and 

 c-ujirn- i-lili>riil-; tin- high reactions with polarization, 

 clirni: ilium .-alleviate, calcium nitrate, uranium 



nitrate. strontium nitrate, copper nitrate, and mercuric 

 chloride ; tlu> miMlerate reactions with iodine, gentian 

 \mlit. -afniinn, HIU! tein|M-rature ; the low reactions with 

 cobalt nitrate nml tximim elilnride reactions ; and the very 

 low with rhloral hydrate and potassium 



sulphide. 



' vl.rid the very high renetioiis with |i 

 gallic aenl. ni!- -ulphnric a> id. hydrochloric acid, 



potajwiiini hydroxide. |>..tn-.-itiin ii-dide, [mta-Miim sul- 

 I'h \nimte, sodium hydroxide, and sodium sulphide ; the 

 high wnh polari/Htion, ehrniiiie acid, MM! in in 



JIM nitrate, uranium nitrate, strontium 

 nitrate, enp|>er nitrate, cupric chloride, and mercuric 

 chloride: the moderate reactions with iodine, gentian 

 VIM], -i, .safranin, and temperature; and the very low reac- 

 tions with chloral hydrate, potassium sulphide, cobalt. 

 nitrate, and barium -blonde. 



lowing is a summary of the reaction-intensities: 



NOTES ON TUB TRIBES. 



Among the very striking features of the four charts 

 are: 



The closeness of all three curves in each chart and 

 the wavering relationship of the hybrid curve to one 

 or the other or both parental curves, occasionally going 

 or below parental extremes in Charts E 30, E 31, 

 and E 33, and frequently (15 out of 26 reactions) in 

 Chart K 32 ; the close correspondence of the curves of 

 the three sets of rhizomatous irids (Charts E 30, 31, 

 and E 32 ) ; and the very definite differentiation of the 



- of the rhizomatous and tuberous series. 



In the first set the cross is between members of the 

 rabgenera Oeocyclut and A pagan; in the second set, 

 between members of the subgenera Ococyclvt and Pogo- 

 nirif and Regelia; in the third set, between members of 

 ubgenus Pogonirit and Rtgelia; and in the fourth 

 ."t. between members of the subgentu Juno. In the 

 three sets of rhizomatous irids the curves are so nearly 

 alike as to suggest that the snbgeneric division of Ha>- 

 sellirin;: referred to in Part II is botanically largely 

 artificial, and that the primary division into rhiznmaton* 

 and tuberous groups is well founded in expressing funda- 

 mental botanical differentiation. Although only one set 

 of tuberous irises wu studied in detail in this research, 

 cursory investigations were made with other members of 

 this series (including /. hixlrio Reichb., /. tingitiana 

 Bows and Rent., /. rtlifvlaia M. Bieb.. I. alata PoTr., and 

 nojrira Hoffm. ; the firrt three belonging to the rob- 



- Xiphion and the last two to the subgenus Juno), 

 in all of which the reactions were in cloae correspondence 

 with those of this set In the previnn research with 

 irid starches it was found that the members of the rhizo- 



8 



matous aeries have in comparison with those of the tuber- 

 ous series, beaidea different ln-tol..-,, |,r..|Ttie, a lower 

 degree of polarization, lower reactivities with iodine, 

 higher rcaetiuthv with gentian violet anil nafranin, and 

 di-tmctly higher tcm|>eraturva of gelatnn < 'wing 



t.. ini|>n)|)er strengths of the reagents, evidence waa not 

 recorded that is satisfactory to differentiate the tarchea 

 then Mudied; hut there was clear c of grouping 



of the two series, the members of the rhizomatnus serial 

 having, as a whole, higher reactivities with chlorn 

 drate and chromic acid, and lower reactivities with 

 chloride and 1'urdy's solution. These results t 

 ai'eurd with tlnw< of the prvw-nt n-w-areh, there U-m^ in 

 the rhizomatouM sericti mean lower rca< ti\itte t . with |Mila- 

 rization and iixlinc, higher n-activitii-s with gentian 

 Molet and nafranin, higher t. ni|-rature of geiatinizatimi. 

 higher reactivity with chloral hydrate, the Bane or a 

 tendency to a higher reactivity with < hnnnic aeid, and a 

 lower reactivity with potassium hydroxide. 



The types of curves of the rhizomatous and tuberous 

 irids, respectively, differ chiefly in the relative townees 

 of the rhizomatous curve in the reactions with pyrogallic 

 aeid, nitric acid, hydrochloric acid, potassium hydroxide, 

 potamium iodide, sodium hydroxide, sodium sulphide, 

 calcium nitrate, uranium nitrate, copper nitrate, tupric 

 chloride, and mercuric chloride, and the highness in thoae 

 with chloral hydrate and sodium oalicylate. I'rohahly 

 among the irids will he found MUIIC -j>e<-ie< or hylirid that 

 will, as in case of the crinums, bridge the two Aerie*. 



Owing to the almost invariable closeness of the three 

 curves in each set, opportunity is rarely afforded for a 

 satisfactory study of the relationships of the hybrid to 

 one or the other or both parent*. It will be seen by the 

 following summary, the figures of which are to be taken 

 as having only tentative values, that the different hy- 

 brids vary in their parental relationships, especially in 

 their intermediate, highest, and lowest records. 



The following i* a summary of the reaction-intensi- 

 ties of the hybrids as regards sameness, inli run ilinlintm, 

 excess, and deficit in relation to the parents: 



The differences in the reactive-intensities of the rhi- 

 zomatous and tuberous series are indicated in the fol- 

 lowing table: 



M: 



.. 



mrim: 



I. ib*rk4fDJuM-hmaU ..... 

 I. iUricB-wwfaltf-Oonk . . . 

 I. Mnialli-(MUlid*-inra. gny 

 TubvixM MTM: 



Vy 



Hich. 



a 



31 

 3J 



8.7 



Mod- 



8.7 



83 

 9.7 



Low. 



7.7 







IS 



0.7 



47 



* 



a.7 

 I.I 



