H HM WWW l 1 WIWWI N l N l l .5} METHOD FOR. DETERWMHG SEED WABIUTV BY ELECTRICAL CC N ENC"; [VHF MEASUREMENTS Thesis for Degree of M; S. GEORGE L. FICK 3 924 \ 2.1.3.1: :an4 ‘ ‘ V THESIS t ‘ ..tho. --I“ ‘I ‘ I. : i.~¢l‘.§fifiufl4flli&hfifiuw1hW, .-.. .. I 13.11.!!! “.41 . Li. I: A LIE’PHUD TOR DEBT;‘JLIITIITG SEED VLLBILITY BY “‘7 7‘ . (th '1‘? d’V'Tu' ”‘1 “rm 3"" (11 ‘7.- ‘1‘7 ‘\ LuduL${I‘JLLJJ UL’l‘quiq‘ [VLLI A..-ALJLRQ..-£A‘ITSQ TiLJSIS Submitted to the Faculty of the Kichigen Agricultural College in partial fulfillment of the requirements ‘Ir for the degree of Taster of Science. By .f{ I .L, )5 é? A TV George L. Fle. 5.... Mull {THESIS I. III. IV. VI. VII. VIII. F3 ABLE CI“ CM x "" L‘~fi7' h” Lihsido I..ITLUJU“”ION. 145.1". BiIilLS {£17 ICZITI- ES A. 7334;111:1110}: OF éL-Ll DIJIIITTJYS B . RATE LlLS . C o L}¢::_.IV{ALL :32?" 1rep 0 1;"..112 .'"";‘S .1771) 1. LJ‘ILTS. I‘Vr- ('9 ('7, -L J4“ L: .J—JJJ‘ km 31:} ..LJ IL) (1:14.131. DEJ J'RIL‘IIOI.’ 1131‘. OF IJL T'T'LJJ 107’ J. I Ctr-'1 & nnil D FOR DETERIIHIKG SEED VlnBILITY BY ELEG' ‘75: IthlL C ()IDUC'UIV I TY 13951213; 3.1738 . I. IKTRODUCTI H. The first use of electrical conductivity in physiological re- search dates back to 1836 when Eduard Veber (l) at Halls studied the elec- trical resistance of the body and the effect upon it of temperature and moisture. Following him, came du Bois Reymond (3) with his studies on the resistance of muscles; Ranks (6) who showed that the resistance of plant and animal tissues decreased upon death; Stewart (4), Roth (5), Burgerszky and Tangl (6) who studied the resistance of red blood cells; and many others who need not be mentioned here but whose work will be found discussed in any good standard textbook. or the numerous recent workers, only a few will be briefly re- ferred to. Osterhout (7—11), through his thorough studies on the resist— ance of living and deed tissues of several marine algae and his conclu— slons based on these studied as an indirect means of measuring permeabil— ity, gave a big imyctus to the use of c nductivitg methods in ghysiologi- cul yroblems. Green and Larson (13) and Johnson and Green (15) in their work on the conductivity of bacterial and yeast cell suspensions respect- ively s owed that, at death of these organisms, a change in resistance took place due to an eltergtion in the permeability of their membranes. Such men as Veshourn (14-15), Taylor and Acree (16), Hibbard and Chap— man (17), and Green (18) deserve mention as havinm meme electrical con- ductivity methods accurate and attractive in the investigation of many physiological phenomena. Following Osterhout's inportant findings in his studies on the permeability of Laminaria and especially the difference in conductivity of living and dead tissue of the same, Dr. E. a. Bessey, of the Department of Botany, Michigan Agricultural Collefle, Conceived the idea that electrical conductance measurements might constitute a means of determining seed via— bility that would be a decided step in advance of the present methods of running germination tests, which require at least several days and, in the case of some grass seeds, several weeks. Promgtly, work was begun by Doc— tors Bessey and Hibbard to investigate the problem suegested by this idea. Then but fairly begun, the war intervened and the work was dropped. The idea, however, persisted and in the fall of l 25, aided by the liberality of the Ferry Seed Company in contributing a fellowship to carry on the work, the investigations were resumed by the author. The problem briefly is this. To determine a correlation, if there be any, between seed viability and electrical conductivity. If a re- lationship between the two could be established, attention would then be directed toward making resistince measurements a practical means of deter- mining the percentage of germination of any sample of seed. It is believed that a correlation has been feund wiich is fundamental and which, with fur— ther nendment and improvement, can be practically used. however, much work remains to be done and it is realized that this is but a preliminary investigation. I I. ILnTER I;-.'.JS “I'D le‘I'ODS . A. Description of apparatus. The apparatus used was practically the same as that recornended by Hibbard and Chapman (1?) with several minor changes. is shtwn by these authors, the apparatus is capable of great precision and yet is easily 0p- erated. The average time required to make a reading is about two minutes. The source of current employed was a 60-cycle rotary converter actuated by the college circuit of 230 volts (D.3.). In connection with the converter, a variable transformer was employed which cut the current down to 6 volts (A. C.). This was then led directly to the bridge ind *alvanom- eter. The converter and transformer were placed under the table at which the work was carried on so as to prevent them from exerting an influence on the sensitive galvanometer. A switch Vas attached to the side of the table to facilitate the starting of the converter. Originally, a radio oscillator was used as a source of current in connection with a telephone tuned to the frequency generated by the oscillator. This, however, was early disearded because of the difficulty of obtaining an accurate "minimum" in measuring high resistances, and because it is impossible to use the telephone success- fully except in an absolutely quiet place such as a telephone booth or in- dividual laboratory. An attempt was made to use the galvanometer as a de— tector with the oscillator but the latter did not generate sufficient current to actuate the gaivanometer coils. Although a high frequency and an ex— tremely pure sine wave such as the oscillator gives are desirable, Green (18) and Hibbard and Chapman (1?) have shown that very good results nay be had by using a frequency of 60 cycles. The bridge used was a Leeds and Northrup product and was of the -4.- Kohlrausch roller type which has proved very satisfactory wherever used. The bridge wire is 470 cm. long. The scale is divided into a thousand di- visions which, in turn, are divided into halves that can be read accurate- ly to a fifth, making the error from bridge readings negligible for all purposes of this work. By removing two plugs, the bridge wire can be ex- tended where extreme accuracy is desired in measuring low resistances. In this work, the "short" bridge was exclusively used. Prior to use, the bridge had been sent in to Leeds and Northrup to be calibrated. The known resistances were of the "plug decade type." The plug; controlled "five decade box" put out by Leeds and Northrup has a range of from.l to 20,000 ohms and an accuracy of 1/20 per cent for all but the 1 one coils (which have an accuracy of 1/10 per cenq. Such a box was used throughout these experiments and proved very satisfactory. Care was taken to keep the plugs and "plug holes" clean at all times. An alternating current galvanometer of the Rowland electro- dynamometer type was employed to determine when a balance had been reached. This instrument, also a product of Leeds and Horthrup, has proved very satisfactory. The stationary coil of the galVanemeter is placed in the main circuit and the swinging coil across the bridge, just as is the tele- phone in the Fohlrausch method. The maximum allowable current of the fixed coil is 1/5 of an aspere for 10 second periodsL—the value just reached by the current from the transformer. The swinging coil has a m“;imum allow- able current of 1/10 ampere for 10 second periods. Alternating current galvanometers of this type must be protected against external electric fields. The main offender in producing such a disturbing factor was the rotary converter until it was placed under the table, as previously stated. A key was inserted in the line running from the bridge to the swinging coil. -5— v This enables the Operator to break tie circuit as soon as a deflection of the galvanometer scale is noticed. After an adjustment of the bridge is made, the key is again pressed down to make contact-~3ust long enough to get a deflection. This is continued ntil a balance is reached. By closing the circuit for an instant only at each trial, the galvanometer is protected and, moreover, errors from heating and polarization are avoided. To avoid too great deflections of the galvanometer, when read- ings were first begun and the difference between the unknown resistance and the known resistance in the box Was likely to be great, a plug resist- ance box was put in the main line to enable the Operator to cut down the current to the stationary coils of the galvanometer. Usually, a resistance of 300 o.ns was inserted until, after two or three trials, balance was approached; then this resistance was taken out to secure greater sensi- tivity for final bridge adjustments. The electrolytic cells ultimately used were of the immersion type (see figure 3). They were made by Eberbach Brothers of Ann Arbor, michigan, according to the designs and specifications furnished them. For the first measurenents made in this work, a U—shaped cell was improvised with electrodes 1 0m. in diameter and 8 cms. apart, but this was soon dis- carded as the immersion cells proved better adapted to the work in hand. Prior to use, the electrolytic cells were thoroughly cleaned and the electrodes platinized according to the nethod recommended by Findlay (19). When ready for use, the Cells were immersed in tie solutions to be measured. The solutions were in Pyrex beakers which, following the addition of the cells, were placed in a constant temperature water bath regulated to 2500. The water bath was essential since temperature changes produce marked differences in resistance readings. Since all readings were run in pairs, two electrolytic cells of exactly the same specifications and made by the same nan were used. These were checked against one another in conductivity water and agreed to within 5/10 of l per cent, which was close enough for this work, es- pecially since this error was compensated for in using first one, then the other, in making readinas on successive samoles of seed. The awparatus is illustrated in figure 1, and a diagram of the "set-up" ultimately used is shown in figure 2. B. Iaterials. Only tno kinds of seed were used in this work as it was thought better to experiment intensively with a few sorts than superficially with many sorts. Timothy aid red clover were chosen as they represent two very common and widely different types of seed--timothy having a very large en- dosperm and red clover none at all. Furthermore, they are of convenient size to work with. The seeds were obtained from various sources. The D. K. Ferry Seed Comoany sugplied 10 pounds of timothy and 10 pounds of red clover from their 1923 crop. A quart jar of 1915 timotLy was obtained for the work from the same source. Attempts were made to locate very old seed of each kind to be used with the highly germinable seed furnished by the Ferry Seed Company in making us mixtures of different percents of germination. The various attempts to procure old Seed proved futile, no seed older than 1918 being found in the time allotted. Some old class exhioits in the Rotany Building were gone through and, finally, a sample of red clover dated 1895 Was uncovered. Ho old timothy was found. The four lots of seed mentioned formed the nucleus of seed used in the experiments of this work. The germination of each lot of seed was very carefully deter- mined. nfter being thoroughly mixed, 10 samples of 100 seeds each were counted out from each lot. These were placed in moist chambers and the percentage germinntion calculated for each sample; then, the average of the ten taken as the percentage germination of that particular lot. Beginn- ing after three days, the seeds began to germinate and every few days the Sprouted seeds were taxen out until, after two weeks, the final counts were made. The germination percentages for the four lots calculated, in this way, follow: Lot #1, 1915 crop timothy (from Ferry Seed Company) --- 73.9% n 3’2, 19.3.5 " H n n n n ___ 89 . 8% " #5. 1935 " red clover " " H H --- 91.8% " #4, 1895 " " " (frou Old Botany Build—)——- 5.6% { ins sample Since no timothy of very low germination could be found, dif- ferent methods of killing these seeds were tried out-—first, by etheriza— tion; second, by chloroform; and, lastly, by heat. The resmlts of these experiments are tabulated below: Table 1. Effect of Etherizetion on Germination. : Average gercent germination (5 series) after ex- posure to saturated ether vapor (under sealed bell Seed : jggj for : 13 hrs.:24 hrs..56 hrs.:dS hrs.:EO hrs.:7 days Lot 1Ll915 timothlj : 81.7 : 78.6 : 77.5 : 74.0 : 77.2 : 70.6 Lot 2L1923 timothyl : - — Table 2. Effect of Chloroform on Gennination. Average gercent germination (3 ser ries) after ex- gosure to saturated chloroforw Vsioor (u;1der bell Seed : jar) for : 2 days : 4 days : 6 days : 8 days Lot 2(1923 timothyl, : 71.4 : 71.2 : ‘2.0 : 63.5 Table 3. Effect of Dry Feat Treatments on Germination. : PL rcent germination after folloving heat treat- Seed : mants : 5 d iys at : 3 days at : 7 days at : 7 days at : cooc. : 9003. : eooo. : 90°C. : dry heat : dry heat : dry heat : dry heat Lot 1(1915 timothyl : 15.0 : Tone : Hone : Hone None 0 L“ H H. ’7‘ ;, Lot 2(1935 timjthy) - - ' Slivht . The conductivity water used throughout this work was triple dis- tilled. The condensed steam from the college power plant was redistilled in a Barnstead automatic still which yielded water having a speC1fic resist- ance of 45,000 ohms. This, in turn, was redistilled in an all-glass still yielding what shall hereafter be referred to as "conductivity water" of a Specific resistance of 35,768 ohms (or Specific conductivity of 7.95x10_6). although not conductivity water in the strictest sense of the word yet it was a very pure and high grade of distilled water and satisfactory for this preliminary investigation. C. Method. Various methods of handling the seeds vere tried out but the method that will be here described is the one finally adopted. The others will be briefly mentioned under the exyeriments discussed farther on. Both kinds of seed were handled in the same general way except for minor dif- ferences, but they will be separately considered. After working vith amounts varying from .5 gm. to 5 gms., the former (used with 100 cc. of conductivity water) was selected as the best. Smaller amounts, while perhaps giving better results, would hardly be fair samples. A .5 gram sample of timothy contains betveen 1600 and 1800 seeds vhich, vhen the seed lots are well mixed, should give a fairly representa- tive sample from the standpoint of germination. a number of .5 gram amounts were carefully weighed out on a standard Becker scale, using weights standardized by the United Stated Bureau of Standards, and stored in small coin envelOpes properly labeled. When ready to run a test, two of the envelopes—~the tests were run in pairs--were Opened and the seeds emptied into separate Pyrex beakers of 125 cc. capacity. The time was carefully noted and, on the minute, 100 cc. of conductivity water was added to one of the beakers and, after a three-minute interval, 100 cc. to the other. This interval was main- tained throughout until the reading was made. Fifteen minutes after ad- ding the conductivity water, the solutions were well stirred for three minutes, an attempt being made to stir both samples uniformly. After another interval (of thirty minutes) the solutions were again stirred~~this time for two minutes. Following this stirring, the electrolytic cells were lowered into their respective solutions and the whole allowed to come to equilibrium before the resistance reading was made. The reading was made exactly one hour after the addition of the conductivity water. One gram of clover was chosen as the best amount to work with. Such an amount contains between 650 and 800 seeds.. A number of samples were weighed out and stored in properly labeled coin envelOpes just as in the case of timothy. Yhen a test was run, two of the envelopes were Opened and their contents emptied into separate Pyrex beakers of 125 cc. capacity. On the minute, lOO cc. of conductivity water was added to one of the beakers and, after a four—minute interval, the same amount to the other. Immediately after the addition of conductivity water, the solu- tion was stirred for two minutes. Fifteen minutes after this addition, the solution was again stirred for two minutes. The solution was stirred for two minutes each time at the end of the two following successive half-hour periods. The cell was introduced into the beaker after the last stirring and, shortly afterwards--exactly one and an Lalf hours after the addition of conductivity water--the reading was made. A clearer idea of the methods followed for timothy and red clo- ver may be gained from the schedules on pages 16 and 34 respectively, which are the forms in which the data for all tests were recorded. III. EZE’SHLZZIT’BS .LITTJ REQULTS. When this vork was first begun, it was thought that the most logical way to attack the problem was to measure the resistance of the seeds themselves, much in the sane manner as Osterhout measured the re- sistance of the Laminaria discs. But, when carried into practice, this method is beset with difficulties. As mentioned by Osterhout, the surface of the discs in contact with the electrodes is a great source of error and cannot be kept constant. He obviated this error by holding the discs off from the electrodes a short distance so that the current traveled from the electrodes through a small chamber of sea water before reaching the discs. With small seed, such as timothy and clover, this is not easily done. It was thought that, if the seeds were placed in the bottom of a small U- shaped cell and the electrodes lowered down the sides of the "U" until just above the seeds, this difficulty would be surmounted. But, now, the difficulty of obtaining uniform packing of the seeds and of keeping some of the seeds from rising upward until they rested against the electrode surface preSented itself. One complexity led to another and attention was turned to discover a better method. Several schemes were tried be- fore the method finally used gas discovered and adoyted. In the followe ing experiments, the results obtained from these preliminary eethods are reported together with the final nethod. Experiment 1. The electrolytic cells used in this work were made according to the same design and specifications yet it is impossible to make cells so alike that they will Wive exactly the same resistance reading for a given solution. 80 an experiment was carried on to check the readings of the cells on hand——four of them-~to determine the two cells that would give closest agreement. For this purpose, several readings Jere taken, with each cell, in conductivity water kept at a constant temperature of to 0 v “‘ ‘ " . 0 1 5 3. The cells had all oeen yPCVlCUSlJ cleaned and the electrodes plat— inised according to the method recomnended by Findlay (19). They were Lanersed in 100 cc. conductivity water contained in Pyrex bankers that had be n carefully cleaned and rinsed with conductivity water. The beak- (D ers containing the conductivity water and cells were placed in the water bath, regulated to 2500., fifteen minutes before tle readings were made. Several readings were made with each cell following the same procedure. The readings obtained were carefully recorded. Comparison of them showed that cells ”1" and "4" gave the closest lyreement and most constant readings. The electrodes of cells "a" and "5" sieved bright areas where the glatinum black had not been degosited due either to impurities in the platinum or to dirt that had not been removed from the electrode in the cleaning. The electrodes of cells "1" and "4" swowed a uniform velvety layer of platinum black and this must have accounted for their greater constancy. They were used in all further experiments. The aver- age resistance of the conductivity water, according to the three separate readings of cell "1", was 135,685 ohms; according to the readings of cell "2", 125,851. The difference between the average reading of the two cells is .13 of 1 per cent, which is the largest possible error that can result from their use. This was deemed allowable in this preliminary work, especially so since relative rather than absolute resistances were sought. Experiment 2. This experiment was performed to determine the resistance of the conductivity water used. The conductivity water as previously men- tioned under "Katerials", was trigle distilled, the last time in an all glass still. The distillate care through the condenser hot enough to rid itself of ammonia vayor. The water was keit in very old five-gallon g ass bottles. After standing in these bottles (tightly corked) for two weeks, the water shoved only .6 of 1 per cent decrease in resistance. The error resulting from such a decrease, in measurements of solutions of 8,000 to 14,500 ohms resistance-—the working limits in this investigation--would be very small and compatible with the aims of this paper, namely, to point out a fundamental relationship. The error mentioned above was further de— creased by making up fresh conductivity water every six or seven days. The following table gives the results of three readings with the two cells adopted. Table 4. Resistance (in ohmsl of )onduetivity Water. Reading : first : Secon : Third : fiverage Cell "1" : 125,709 : 135,649 : 125,696 : 125,685 Cell "4" : 135,974 : 125,783 : 125,798 : 125,861 Average reading (00th 09115) = 125,768. Conductivity - ' l or 7.95 x 10—6 125768 E:perixne11t 5. After obuidoninq the method of Lemsurinx the resistance of the seeds themselves, the matter of relative absorption and excretion of SultS, as indicated by conductivity veesurements, was considered ns a possible means of determining seed viability. Five-tenth gram amounts of ti othy were soaked in HsCL and Eli solutions of varying Concentration for differ- ent lengths of time. The dry seeds were placed in an ingrovised sieve mede out of a 10 cm. length of lnrae glass tubing (4 cm. in dieneter) with H, a pcrtition o paraffin across one end of it thut was perfornted with holes small enough to keep the seeds from gdssinfi through. The sieve, contain- ing the seeds, was introduced into a beaker contuining n definite amount of ‘the salt solution. lfter standing in this n certain lenrth of time, the seeds were washed by pouring n wiven amount of conductivity water through the sieve to renove the salt from the surface of the seeds. The outside of the sieve was also thoroughly cleansed of the Snlt. The sieve plus the washed seeds vss then pieced in a beaker containing 70 cc. of conductivity voter end the electrolytic cell (i nersion type) witiin the sieve. After a short interval, a resistance reading yes wide and, thereafter, at reg- ular intervals. The readings measured the resistance of the conductivity. water plus the salts excreted by the seeds plus the seeds themselves. This method was unsatisfactory because (1) it gave very erratic results, (d) reuuired too much time, and (3) because it was difficult to perfect a techniine whereby each lot of seed would be given exactly the smne treatment. Loreover, in some cases, almost all of the seeds rose to the surface of the salt solution and it was imgossible to submerge them. In other cases, very few would rise to tie surface. This inconstnncy, no ocuot, contriouted tcvnrd t’e errntic results obtained which were further — ill-II! augmented by the great amount of manipulation. Although very little work was done with this method because of its combiiCnted nature and because it reguired too much time, it is not imprebuble that a technique could be W develoded that vould pike it ensier ns veil as snorte (L) r. It mus deemed advisable, however, to investigate some other methods. Experiment 4. gnother wethod of nttncninm the groolem from an entirely differ- ent angle was next tried and will be described in this experiment. It is the method Which, With the modifications successively described herein, wes finally adopted. The relative outward diffusion of electrolytes from dead end livinn seeds is nude the basis of determining seed viability. The steps in the develogment of this method are taken up under the folloving subdivis— ions: A. The first vork with this method was done as follows: .5 gram snmgles of seed were put in clean Pyrex bonkers of 125 cc. capacity and 100 cc. of conductivity voter added. nfter on he f hour, reedinns were bezun end, thereafter, at in erVuls. These readings regresent the resistance (in ohms) of the conductivity -v:.lter plus the soluble selts that pass from the seeds out into the water plus the seeds themselves. For the sake of brevity, this resistance is hereafter ref rred to us "Solution Resistance." It is in— teresting to note the gradual droy in resistance of the two lots of seeds recorded in the table. Lot 1 is 73.9 per cent viable timothy; lot 2, 89.8 $7 per cent viable timothy. The engeriment was started at 2:05 .n. Table 5. Solution Resistonces (in ohms) of 2 lots of tinntdy at Various tire intenvdls. Q§_o. 1 2:55 : 5:05 : 5:55 ; 4:05 : 6:05 : 7:30 : 9:15 : 10:15 : 9:o0(nent) 3 1?.11. : i’.L . : 1°.1T. : :P.I;. : 1’.1T. : :P.l:. : .P.l . : 1?.1T. : zl.}I.( dLfi] ) Lot 1 :19076:: 16831 : 15919 : 15575 : 2316:1137? :10856 :10529 : 8484 Lot 2 :19291 : 17870 : 17551 : 15530 : 14160:15$53 :13844 :13359 : 8650 From the preceding table, it can be seen that the drop in resist- ance is stesdJ throu hunt the c;urse of tie experiment but that, at no point, yes there an; marked difference betveen the resistance of tie two lots, and that it teak considerable time for the resistance to droo from 19076 to 8484 ohms, viz. 19 hours. B. Diffusion of salts i1 Water is a comgurntively slov grocess. It was thought that if stirring were employed it would not only bring about the droo in resistance in u slorter period but it night else accentuate t3 differences in resistsnces between tTe tvo lots of seed at any plrtlc- (D ular time. The effect of stirring is brougtt out i1 table 6. The method used N;S exactly the same as that described under "A" except that, in- medietely after the odwition of 100 cc. of conductivity water, the whole vus vigorously stirred with a gloss rod for five minutes. At the end of this time only e few seeds remained on the surface, thus ooviating another undesirable feature of the previous method, for it is evident that all seeds should be submerged if accureCy is to be obtained. The exper'ment was started at 10:15 A.K. Table 6. Influence of Stirring on Solution Resistances Lin ohms) of Timothy it Various Time ntervuls. 2500. LL. 1.-. 3 g. o 2:13 : 3:13 H ...; ()3 : 10:46 : 11:13 : 11:45 : 13:15 Let 1 : 9559 : 8628 : 7761 : 7176 : 6844 : 665 : 6410 11776 : 11462 9980 9105 : 8570 L-i O n lo H C .15 "23 'Q H ' x CD (D ()3 O O A comgerisod f tables 5 and 6 Shove that, where tte solution of seeds and conductivity water is stirred, the resistance dr0ps at a much faster rate and thnt, nt no; one time, there is a nreeter difference be— tween the two lots of seeds. The greatest difference occurs at 10:43 or after the seeds have been in the enter one-half hour. Almost as great o difference, Poiever, is observed ufter tie one—hour intervel (at 11:45). It vould a JGJF that an hilf hear nus too siort s time to expect an ac— curste difference to stov itself in tne relative rate of outward diffusion or leeching of electrolytes from the to; sumyles of seeds esgecinlly since the stirring is done only it the oeeinninf of the test-~nfter that nuturnl diffusion is the only force 09erstive in the distribution of the electro- lyte through the solution. Consequently, it was decided to leave the seeds in the conductivit; water for one hour and to stir tie solution just prior to reading, allowing sufficient time, however, for equilibrium to be established. C. The followinr tvo scleduies, taken from the record sleets kept in tlis ver}, will sfiov in the clearest ninner just how and vhen the stirring 7ns done in the method finully adopted. This is but a slight modification of the method of "B". the tests, it will be seen, were run in pairs. Schedule 1. 2:u8 - .5 an. Lot 1 out in beaker :2z4l - .5 an. Lot 2 put in beaker L,. with 100 CC. conductivity fia-z with 100 cc. conductivity ter : water 2:63 — Solution stirred 3 uinrtes :3z56 - Solution stirred 3 minutes 3:33 — Solution stirred 2 minutes :5:26 - Solution stirred 2 minutes 3:38 - Electrolytic cell immersed :Szfil - Electrolytic cell in beaker ill soifiltiCfll : ”:38 - Reading: :3:41 - Reading: Solution Resistancez“%49 ohms; Solution Resistancezlzefi ohms Schedule 2 will be found on next 9&990 Schedule 2. 3:57 - .5 gm. Lot 1 put in beaker :d:OO — .5 3n. Lot 2 put in beaker with 100 cc. conductivity : with loo cc. Conductivity water : water 7 r-v 4:12 — Solution stirred 5 minutes :4:15 - Solution stirred 0 minutes 4:42 - Solution stirred 2 minutes :4:45 - Solution stirred 2 minutes 0. ,p. (T: O I 4:47 - Electrolytic cell immersed Electrolytic cell immersed in solution : in solution C. c. 0 4:57 - Reading: - Reading: Solution Resistance:9696 ohms : Solution Resistance212700 ohms .0 Other readings were made with the two lots of timothy to de- termine the constancy of the readings. The readings for "Lot 1" varied from 9400 to 10551 ohms; for "Lot 2”, from 2357 to 13852 ohms. nt first glance, these variations abuser too large to be satisfactory but vlen one considers only the tvo most imgortent sources of the veristions——stirring and the taking of seed sumoles--ttey do not seem so. The stirring was all d;.e by bond and all solutions were stirred as nearly alike as it was hu- manly gossible to do, yet it is not unlikely thut one solution was stirred a little more vigorously or, gerhsps, for a little longer period than another and that, in this way, vsriutions in readings that should agree were produced. The diff10u1ty of securing two semqles of seed, from the sane lot, havinn exactly the same germination percentage, Con be readily aggreciuted; also, the vsriutions in resistance readings that would be pro— duced by semoles not hevinm the some germination percentage. In the case of lot 1, the difference between the largest and smallest resistance read— ings obtained amounts to 9 per cent of the smellest reading; in the case of lot 2, to 4 per cent. These variations can be largely attributed to the V "personal error" of the experiment, which with the improvement of the method and the introduction of mechanical stirring devices should be materially decreased. performed to determine D. This part of the experiment was whether the drop in resistance was caused by the leaching of electrolytes from the interior structure of the seeds, or nerely by sone mineral salts adhering to the surface of the seeds. The latter was thought possible, because of the extremely quick drop in resistance but ihprobable b cause of the constancy, within certain limits, of the results obtained. Yet it seemed a point requiring proof. The method followed ‘.':.S the Sane as in "J" except that the seeds vere flushed, prior to use, in order to renove an; electrolyte on the surface of the seeds. The seeds were placed in a beaker with 50 cc. of conductivity water dud vigorously stirred for one minute to give them a thorough washing. They were then filtered off on ash-free filter paper and again vashed by pouring 50 cc. of conductivity water over them. The bottom of the filter paper was functured end the seeds washed into a CUbiC beaker with a few/gentinyunuSJK Conductivity water. The solution in the beaner Was msde up to its full amount by adding what remained of the 100 cc. of conductivity water, and tie experiment contiiued as outlined under "0." The following results were obtained. Table 7. Solution Resistance (in ohms? of Timothy Seeds at end of 1 hour and 6 hour intervals (after a preliminary washing) . 2500. Tile iznnrsed : Lot 1 (75.941 : Lot 2 _L89.8{) 1 hour ; ld,§16 ; 23,442 6 hours : 10,732 ; 14,122 Referring to the schedules under "0", it is seen that the some general relationship of "solution resistances" between lots is maintained in the above table; 9.3., in schedule 1 under "0" the resistance for lot 1 is 9849 ohms and for lot 2, 12,838 cues; in the preceding table, the resistance for lot 1 is 14,612 ohms and for lot 2, 28442 ohms. It is also apparent, from the resistance reading at the end of the 6-hour in— terval, that the fall in resistance is not so rapid as it is when no pre— liminary washing takes place (compare tables 6 and 7). This is probably due to the loss of salts during the vushing period. Eager in ont 5 . Three ways of artificially killing seeds were tested-~otheri- zation, ciloroforming, and dry heat. The best results were obtained by heating in an ovei for se en days at 90°C. in the case of timothy. Two lots of timothy were Killed in this manner—-1ot 1 (75.9% viable) and lot 2 (89.83 viable). Both lots were put in and taken out of the oven at the Same time, receiving the same handling in every way. It was intended to use these artificially killed seeds of both lots in nuking up m’xtures of various germination percentages from O to 89.83, but before using them this present experiment was undertaken to compare tiem from the standpoint of relative resistance. The method outlined under "3" of Experiment 4 was used. All readings were made at the end of one hour and at a temperature o of 25 0. Table 8. Solution Resistances (in ohms) of Artificially Filled Seeds. Lot Kumber : Test Number : 1 : 2 : 5 : 4 l : 8,983 : 8,643 : 8,870 : 8,425 2 : 11,983 : 11,783 : 11,883 : 11,8”2 The table shows that the solutions from the two lots of "dead" seeds offer vastly different resistances to the passage of an electric current. It 5 one, further, that seeds artificlaily killed by dry heat cannot be used in making ug samples of different percentaxe germination. Experiment 6. Tie method of procedure described under "C" in Experiment 4 gave such constant readings after many trials that it was adopted as the best cethod found in the course of these investigations. In this experi— ment, it was used in measuring the solution resistances of .5 gm. samples of timothy of viability varying from 73.9 per cent to 89.8 per cent by increments of 5 per cent--i.e., 75. per cent, 75 per cent, 80 per cent, 85 per cent and 89.8 per cent. Since the results Jf Experiment 5 slowed that it was not Tossihle to use seeds artificially killed oy heat and, since no very old timcthy of low germination could be found, it was im~ possible to extend the limits of this range in germination. The samples mentioned aoove vere made up by mixing the seeds of lot 1 (73.9% viable) and those of lot 2 (89.8fi viaole) in vary‘ng pro- portions oy weight. The weight of each lot of seed to he used in getting a Sample of a certain percent germination was calculated accordin: to the formula X(73o9)+-.5—x{89.8)=.5(y‘, where 5 equals the grams of 73.9 per cent seed to be used and z equals the percentage germination desired in the mixture. For BI&Myle, to find out how many grams of 73.9 per cent and 89.8 per cent seed must be mixed to secure .5 gm. sample of 85 per cent germination, one would write the equation li(73.9)+o5-1{(89.8)=o5(85) Solving, we find x=.151 or the number of grams of seed of lot one (73.9%) that must be used. Then, .5—.151=.349, the number of grams of seed of lot 2 (89.8fl) to be used. By mixin” these amounts of lot 1 and lot 2, a samgle of 85 per cent germination is secured. Three samples of each percentage germination were carefully made up according to this formu- la. The solution resistances of the samgles of various per cent germination were measured according to the method outlined under "0" of Engeriment 4. All samgles vere stirred as uniformly as possible and, to insure this, the strokes made with stirring rod were timed. F strokes were made per quarter the direction of stirring yes in table 9. orty-five minute and, at the end of each such period, reversed. The results obtained are tabulated Table 9. Solution Resistances of Tirothy. 25°C. : Test : Per cent germination : 1 : 3 : 3 :nverage 73.9 : 10,351 : 9,400 : 9, 85 : 9,j72 75.0 : 10,080 : 9,930 : 10,070 : 10,027 80.0 : 10,61 : 10,990 : 11,13; : 10,912 35.0 : 11,211 : 11,627 : 11,(€3 : 11,900 89.8 : 12,800 : 13,853 : ;,337 : 12,590 The results of the figure 4. Experiment 7. above taQIe are graghicnlly represented in The resistance of a solution is greatly affected by any change v in temgerature. Although all of the readings of this Jerk were made at a temperature of 2500., a constant temgerature bath being employed that kept the temperature to Jithin 1/10 of a degree of that point, it was con- sidered worthwhile to compute the temperature coefficient of the solution consisting of timothy seed in conductivity water. Four readings were taken at each of three temperatures--200, ,o , .o. . u . , , . 20 , and 30 Cent1graee--in_tne tvo lots 0: timothy. The average of each four readings was taken as the basis for further calculation. The followu in table gives the readings tenen, using the method described in "C" 09 of Experiment 4. Table 10. Solution Res:istnnces (in ohms) of Tiwothy Seeds at 20°, 25°, and 30 Centigrade. : 2000. : 2500. : 30°C. : 11,284 : 9,849 : 8,568 : 11,086 : 9,095 : 8,474___ Lot 1 = = = (73.9%) : 11,109 : 9,}85 : 0,453 : 11,097 : 9,719 : 3,730 : iveruge=il,114 : ver.ve—- 682 : Lvereve:8,525 : 13,183 : 12,638 : 11,234 : 13,902 : 1,700 : 11,016 Lot 8 : : : (89.82) : 13,891 : 12,800 : 11,170 : 15,844 : 12,827 : 11,224 : Averene-13,855 : Avern3e=12,691 : Averase:11,161 In the case of lot 1, the resistance at 250 is 15.6 mgr cent higher than at 30°; at 20°, it is 15 .1 per cent lighe r th.n st 25°. In 1 the case of lot 2, tne resistnnce at 250 is 13.7 per cent higher than at 30°; at 20°, it is 9.2 1 ° 1 increase for a five—degree droo in temperature is 12.9 per cent, or the average percentage increase per degree drop is 2.6. Suggose a reading is made 71en t1.e te persture is 27°C. and it is desired to knov vhet the reading for the same samnle would be at 25°C. I c ‘1, . - .. . Ber degree droo 18 3.89, the,percentsge incre; der cent higher than at 25 . Tne wver ge pircentnge 88 ‘ o ‘- ‘ . u q ‘ ".1." ‘1‘ _‘ v ." rf‘U—l “ “7' 'u H .'§‘,.', f ‘ " t‘\ L Since the averawe increase/lion ~7 to so 3. vouln 0e 212.8 or 5.4, the factor by which the reading st 370must be increased to give the reading .- 7. , . - . of the 8118 lot at But. Dr stated algeor31Cs11y, 3,7)52,7o+.026(;7-25)2,70 A more general expression of the interpolation of u reuding at any t opera— ture--vithin the limits of 20° — JOOU.--to u temgeruture of ab o. could be t o r-o RL5Q=RXQ+OOLC(X -;b )Rve =R‘.'0+( o UBGX" o 65)R"0 =R"o(.o5+.0sox)* Lttention was nov turned to applying the method, adopted as the best in the Csse of timothy, to another type of seed, insiely, red clover. The following experiments deal with the efforts nude in this direction. Experiment 8. The first thing to deter ine in noplying the hethod to red clo— ver vus the best amount of seed to use add, also, the length of tire the seed should be elloved to rennin in the conductivity vuter before making a reading. This eagerinent deals vith these tvo here or less interlocking phases. It ves decided, frou the first, to use "stirrinp" since it hastens the even distribution of electrolytes thrcuwhout the solution. A. Seeds used were sumsles of lot 5 (91.81 viable) end lot 4 (0.63 viudle). dcheoule 6 (Cl next hire) outlines the procedure and gives the results outline-.3 at certgin tine illtlfl'VitlS. .‘.ll I‘eed'i11:fs were a 0 made at .45 e‘. ‘This formula does not rive absolute accuracy nor is twiS claimed for it. It is liciuded here t; show that the method of seed viability deter inution herein given could be used even though n) Leeis for keepinc the tem- perature constant were at hand.(3ee Discussion of Results) Schedule 5. 9:50 - .5 9n. Lot 5 put in beaker : 9:55 — .5 an. Lot 4 gut in beaker with 100 cc. conductivity : with 100 cc. conductivity water : water 9:45 - Solution stirred 2 minutes : 9:48 - Solution stirred 2 Linutes 10:15 - Solution stirred 2 minutes :10:18 - Solution stirred 2 minutes and electrolytic cell im- : and electrolytic cell im- mersed : mersed 10:50 - Reading: Solution Resistonce:56,655 ohms 10:33 - Reedinfl: Solution Resistance-58,557 ohms {W 10:45 - Solution stirred_2 uinutes :10:4S - Solution stirred 2 hinutes 11:05 - Reading: Solution Resistunee=22,706 ohms 11:00 - Reading: Solution Resistunce=4011db ohms 11:18 - Solution stirred 2 minutes 11:15 - Solution stirred 2 minutes 11:50 - Readingz‘ 11:55 - Reading: Solution Resistance:26,939 ohms : Solution Resistance=15,755 ohms 11:45 - Solution stirred 2 minutes 11:48 — Solution stirred 2 minutes 12:05 - Reading: Solution R sisteneezii749 ohms 12:00 - Reading: Solution Resistune§;19,187 ohms 1:45 - Solution stirred 2 minutes 1:48 - Solution stirred 2 minutes 2:05 - Reading: Solution Resistonce:4,076 ohms. 2:00 — Reading: Solution Resistence:10,770 ohms The same procedure was again followed using .5 pm. enmples of we r e the same seeds but this time the first two stirrings/of three minutes duration. At the end of the first hour, the solution resistance of Lot 5 was 46,512 owns: of Lot 4, 28,445 ohms. At the end of two hours, the solution resistance of Lot 5 was 21,726 ohms; of Lot 4, 10,850 ohms. The results were similer to those in the preceding table except that the readings were somewhat lower out the drOp was proportionally the same —25- during the first two hours. B. Some tests were run, using 1 grmn staples and taking reed- ings at the end of the one end one and an half hour periods. The solutions were stirred every fifteen minutes (for two minutes), and immediately after the lust stirring—-fifteen minutes before the reading was tuken——the elec— trolytic cell was immersed in the solution. All readings were made at 25°C. The following table shows the results obtained. Table 11. Solution Resistances of Red Clover for 1 hr. and l—lZ2 hrs. geriods. 25°C. : Let s (ens-:31 Lot 443.651.) Resistance (in olmsS) :Lot 3 :Lot 4 ::Lot 3 : Lot 4 ::Lot 3 :Lot 4— at end of : : :: : :: : 1 hour : 27633 : 18162 :: — - : - — :: — - : - — 1-1/2 hours ; 17473 : sass ::1602l : 9739 :: 18879 : 9218 The readings at the end of the 1—1/2 hour period were at a better "working level" then at the end of the 1 hour period. In other words, the resistance values were at u level low enough that slight ris~ tukes in weighing the Sungles or in measuring the conductivity water did not produce too lurge errors, yet they vere high enough thnt on aggre- cisble difference existed between the readings for the txo lots. The agree— ment between the readings for somoles of the some lot is not so close as desired, hovever. C. It was decided to use 1 gram sum 1es of red clover seeds end to allow them to remain in the conductivity water for 1-1/2 hours before reading the resistance, then to modify the time and interval of stirring to find out which would give the most constant results. The de— tails of this work need not be given, but the method finally adopted is made clear, and some results obtained oy its use are recorded in the following schedule. (Schedule 4 on next page) Schedule 4. 9:50 - 1 gram Lot 3 put in beaker : 8:54 — 1 gram Lot 4 put in beaker with 100 cc. conductivity : with 100 cc. con‘uctivity ~mater : 'vater 8:45 - Solution stirred 2 minutes 8:49 — Solution stirred 2 minutes 9:15 - Solution stirred 2 minutes : 9:19 - Solution stirred 2 minutes 9:45 - Solution stirred 2 Linutes : 9:49 - Solution stirred 2 minutes and electrolytic cell im— : and electrolytic cell im— mersed mersed 10:00 — Reading: Solution Resistance;l7,695 ohms 10:04 - Reading: Solution Resistance:9,275 ohms Other readings obtained, using precisely the sums nethod, were: for lot 5, 17,551 ohms, 17,990 ohms; for lot 4, 9,182 ohms, 8,689 ohms. This method have, 0y far, the most constant readings obtained with red clover. Some variation in the readings, of con se, cannot be avoided since there are bound to be some samples that are not representa- tive of the germination of the lot of seed from which taken. Exeeriment 9. This experiment deals with the application of the method of "C" under Experiment 8 to samples of red clover ranging in germination from 3.6 oer cent to 91.8 per cent. The amount of each lot to use in a sample of,say, 40 per cent ge minution was calculated from the formu- la x(3.6)+l.-x(91.8)=l.(40), which is identiCnlly the Same as that used in making up the timothy samples. The method followed is exactly the same as that given under "0" of Experiment 8. The following table records the results that Here obtained. (See table 12 on next page) -97- Table 12. Solution Resistances (in ohms) of Red Clover a OFO 508d. Ml) Co u Test . iercentage : : nverag Germination : l : 2 : 5 : 4 . Reading 3.6 : 9173 : 9183 : 8056 : 885: : 8770 10.0 : 8417 : 8200 : 9007 : _Z995 : 839§_ 20.0 : 8989 : 8618 : 9441 : 8208 : 8814 50.0 : 9979 : 10504 : 9423 : 8000 : 9478 40.0 : 9557 : “115 : 11415 : S770 : 968$ 50.0 : 10976 : 10853 : 10504 : 9126 : 10560 60.0 : 9500 : 9619 : 12120 : 982 : 102 5 70.0 : 13776 : 11456 : 12625 : 10351 : 11767 70.0 : 11514 : 1195i : 15585 : 106:1 : 11855 90.0 : 14056 : 17506 : 14906 : 1254‘ : 14254 14500 : 14529 : 15291 : 14587 ' D H O (D ...—a C“. ,p; 10 \1 The above table s one that there is an increase in resistance as the Rermi ation becomes greater and, but for tvo exceptions, the increase is fairly regular. The data included in Table 12 are 3 are graglically in figure 5. In the past ten years, muc. work has been done dealing with the determination of life or death by means of electrical canductivity measure— ments. Eany different giant and animal tissues vere subjected to such electrical determinations and, in sone cases, with noteworthy results. A thorough review of the literature available failed to reveal a sinsle in- stance, h;wever, in which seeds were the suoject of investivation. AS \_4 far as is known by the author, this is the first work in which it was attemsted to determine seed viability by electrical conductivity measure— ments. inlrilv The two methods prelimi—/0d365khave been fully described and discussed under "Experiments and.Results." Since they have been discarded, they need not be again discussed. Tge discussion is limited, then, to the finally adogted method and to the results obtained by its use. In table 9, are shown the results obtained with timothy of various percentage germination at a temperature of 2500. It is readily seen that there is a grogressive increase in solution resistc.ce with viability increase. However, there are some discrepancies. Several rec.— ings for a certain percentage viabilitv are too high. Tris, however, is to be expected as will be explained in the consideration of a specific case. In one test of 3.9 per cent seed, a solution resistance of 10,551 ohms was found, which is not only higher than either of the other read— ian obtained for seeds of that viability but also higher than any read- ing for 75 per cent seed, and almost as high as the readings for 80 per cent seed. This can only be explained by attributing it to a personal error made in technique, or by supposing that the particular sample of seed, giving the higher reading, was of hither fermination than the averaee sample of that class. The latter is a very valid supposition, for, when the germination of the lots of seed were originally determined by sprouting, it was fgund that the different samples used would vary wer cent for lot 3 to 32 yer cent for lot 1. The L. ‘- over a range of 7 germination of lot 1, according to the average of ten tests of 100 seeds o ch, was 75.9 oer cent; the highest test being 83 per cent, the lowest, V“ 61 per cent. This difference was obtained desgite the very thorough mixing of seed lots prior to samgling. So it does not seem unreasonable to suppose that the samgle giving the reading 10,531 ohms was of germina- tion considerably higher than 75.9 per cent. Kore conclusive proof would have been offered in support of the correlation between electrical resistance and seed viability had it been possible to extend the range of germination percentages of the samgles of timothy in table 9. It was originally intended to use artificially killed seeds and, by mixing these with seeds of lot 2 (89.8%), to make up .5 gram samples ranging in viability from O to 89.8 per cent. But, as the data in table 8 show, artificially killed timothy seeds from two different lots did not give the sane resistance. Yet, the readings for either lot are so surprisincly constant that errors in method or technique could not be blamed. Neither lot of killed seeds showed the least Sign of gerninating when tested and were believed to be dead. Reine dead, accordingly, they should show very nearly the same resistance. Since they did not, the con- clusions reached were that either the seeds were not killed (but brought into dormancy) or that the changes that take elace in the natural death of a seed were not affected by tle artificial death produced. Or it may be that, because of the suddenness of artificial death, the seeds retain part of the chemical c nstitution of live seeds which, however, is lost in the longer process of naturally dying. Since the readings for the killed seeds did not agree, it was thought best to dismiss both artificially killed lots and to use only seeds that had lost their vitality in a natural manner. Very old timothy seeds were sought in the hepes of getting some of very low germination. fihen none was to be had, the tests were confined to the limits represented by the germination of lots 1 and 2. An interesting sideline of the resistance measurements, taken ~50— With timothy, 7as the determination of What produced the conductivity of the solution. After a test had been run, using a lot 1 sample, the solution was filtered to remove the seeds, and a portion of the fil- trate sent to the Experiment Station Chemist for analysis. His report showed that 14 per cent of the mineral content of the solution was P.O ': H 5’ that an appreciable amount of Ca was present (probably as CaHPO4), and that there were traces of 304, fig, and K present. Due to the minute quantity of miner l matter present in the solution, he succeeded in making a tive fuantita- test for Pgow only; for the others, only qualitative tests 0 '7e re mad e . The most important point brought out by the data is that a difference does exist between the solution resistances of seeds of high and low germination (see figures 4 and 5). That is the fundamental re- lationship that must form the basis of any future work that looks toward the practical agplication of this method in the determination of the ger- mination percentage of any lot of seed. The same method was used in measuring the solution resistance of red clever as mas used for timothy except that 1 gram of the former was used and that the time and interval of stirring were slightly changed (see schedule 4). The readings obtained for samgles of red clover, ransing in germination from 3.6 per cent to 91.8 per cent are shown in table 12. A glance at the dais of table 12 suffices to tell one that the resistance increases with the germination but, ujon analysis, one finds several discrepancies. Some readings seem too high, others too low. No more can be said regarding these than has already been said concernin, those in the table of timothy readings, viz., that they re51lted (1) from a sample having a much hisher or lower germination than it was -3l— supposed to have, or (2) iron a "personal error" in conducting the ex- periment. The average readings for red clover lien incorporated into a graph (figure 5) show that, vith t;e eaceytions, the resistance increase is concomitant with viability increase. That is the important thing for it forms a foundation uyon which future work Can be based that will make the method practical for determining the viability of red clover. The immediate object of this investigation was not to develop a method whereby seed viability could be measured by electrical conductivity, but to attempt to discover a fundaaental correlation betweex seed viability and electrical conductivity. Once established, such a basic relationship night be put (D to greeticai use in the laboratories of 3 ed firms, as a snicker and cheaper means of determining the bercentage of germination of any lot of seed, than the present gethod of running germination tests. The tesgerature coefficient of resistance was worked out for timothy merely to show that it is not absolutely essential to have such an exoensive piece of agpiratus as a constant temperature Hater bath to conduct the tests; also, to show one way in woich a temperature coefficient might be calCulated. Vhen a practical method, based on the correlation discussed, is perfected, the work of developing it will have been done in a laboratory where a constant temperature bath is available. All the readings, for a certain Kind of seed, will have been taken, then, at the sane temperature (9.3., 2500.) and embodied in a able or graph. If a temeerature coefficient be carefully worked out at the save time, a vorker in another laboratory could make his test at room temperature and, by means of the temperature coefficient, interpolate the resistance reading value. Referring to the table or graph containing LI ..a ('1’ U) l t (I! C 0 obtained t' 1 1 tne solution resistance values for se:ds 0: various percentage germina— 0 . ‘0 1 I ~ 0 0 V ' ‘ ‘ ‘ tion at do 0., he could find unnediately tne Viability of that particular sanple. The coefficient will vary in value for seed of different viability but this could, no dcubt, be mathematically calculated and included in the formula. Kany flavs can be found in the method described in this paper and the writer is fully cognizant of them. Following are some of the recommendations for future work that will tend toward the elimination of these faults. 1. In this investigation, all the stirring was done by hand. as previously explained, attention was given to stirring each sample ust as nearly like the others as was humanly possible. To stir all exactly alike, hovever, is impossible. Likewise, it is impossible for tvo uen to stir exactly alike. To obviate the unavoidable errors result— ing, it would seem a great improvement to stir the solutions mechanically. This would not only give more constant and dependable results but, it is believed, it vould also make the method shorter and easier to conduct. It has been shown that the drop in resistance of a solution to the "vorxing level" occurs much more Quickly vhen the solution is stirred than when not stirred. If the solution were stirred continaally, by mechaniCal neans, there is no doubt that the drop to the "working level" would require yet less time. 2. It would be advisable to use larger samgles of seed and probortionally more conductivity water. The difficulty of getting fair samples, from the standpoint of germination percentage, has been emphasized and it is logical to suooose that the larger tie sample the more nearly it will agproximate the fiermination of the veil-mixed lot from which it is taken. 3. The solution resistance of timothy in conductivit' water was taxes at the end of one hour, i.e., the readings were begun at that time. The average time required to take a reading was tvo minutes. Fovb ever, in one case it may have required less than tvo minutes, in another case, longer. if something occurred that delayed the reading (and it did several times), the solution resistance would be low due to the outward diffusion from the seed over a longer period. It mould seem an advantage to draw off a definite (mount of the solution just as the one-hour period (or whatever it may be) elapsed. Then, it would matter little vhether the reading vere made inmediately or shortly aftervards. 4. The conductivity vater used in this work vas made up fresh every six or seven days. It was always prepared in precisely the same manner and by using the sane distilling separatus. Successive lots agreed so closely in resistance that the difference betveen then was ignored in calculating the solution resistances, since relative rather than absolute results vere sought. But, if the method were used in different laboratories, it would be extremely difficalt for all to make or secure ccnductivity water of the same resistance. A means of calcalating the solution resistance is needed that takes into consideration the resistance of the conductivity water used. Which will enable work rs in various laboratories to use con- ductivity water of their own making and, for the same lot of seed, to ob- tain identical results. 5. Very satisfactory results vore obtained vith the electroly- tic cells used but there are several reasons wk; on ogen type of cup, such is used with the soils oridge, vould prove more satisfactory. In the ,_ U1 r‘,‘ ‘1. -‘)~:. first glass, the cells are ver} frafiile. secsuse the electrodes are L g. 1 enclosed in glass (see figure 5), trey are very difficult to clean yrior to reylatinising or to keep clean. furthermore, they are difficolt to ineke so similar that any tvo cells Jill give the same resistance reading for a given solution. An been cup made of .erd rubber would seem much more desi able. It would be chesger, less fragile and easier to keep clean. Receuse of its oyen nature, it would not be difficult to make the electrodes J of exactly the sews dimensions and the same distance apart. 6. Eur future vork, it is essentisl that old symples of timothy and red clover seeds be ,rocured that will be of oil varying degrees of via- bility. The solution resistance or a swigle or red clover thst is, ye shall assume, 50 per cent vinole could then as Co pared Jith that obtained for the 50 per cent samples used in this lurk, lilCh vere made up by nixins red clover seeds o: lots 3 and 4. In other words, the solution resistances obtained r-mixed squies Could be compared vith natural sauyles having 1 0! for artificiali he same germination >ercenteges. This seems a very immortant consideration. 7. Seeds from various sections of the-country, having the some viability should also be secured to see what influence soil hns on the elec— trolyte content or composition of seed as inuicated by comparative solution resistance measurements. 8. It is possiale to conceive n SHHNIC of, say, 50 per cent germination riving the smie solution resistance as a 70 Per cent samyle. The seeds of doth sufiples might Contain a like number of dead cells (having permeable membranes) but due to the nore fortuitous IUCLLiifl of such dead cells (farther renoved from the germ) in the one samble, a higher viabil- 1.1lifht 7 ~ 0 ‘ - ‘ . n o - o _ o ity/ootsin, ultheufh o: s Hexner nature. This posSLOle relationship needs to be investigated. (‘v-I r‘ H n a n v. L) 14"O-IL£“I»“Y. These investigations dealt with an attempt to determine seed viability by an electrical conductivity method. The seeds used were timothy and red clover. Three different methods were used in studying the possible correlation betveen seed viability and electrical resistance. These 'vere: l. leusnring the resistance of the seed; themselves. s. Comperinz the relative absorgtion and excretion of salts. u. Keasuring the relative c.tverd diffusion of electrolytes as indicated by conductivity readings. :ods one and tJO xere discarded. l ethod three "lips adoihted. The readings Detained of its use seep t3 inticute that a correlation between seed viability and electrical conductivity exists which, with future improvement, may be capable of practical ayplica- tion. V I o -XCIETO'VJLEGTE‘IT o The writer vishes to express his deep aggreci- etion to Doctors R. P. Viobdrd and E. A. Ressey for the many kind sunyestions and the very helpful criti- cism that they gave him cancerninu his work; also to Doctor Robinson for much help in connection vith the aiparntus used, and to “ref. 3. I. Fuskell for his valuable assistance in degling vith the mathematical aspects of the froblem. “dual- \ f, «aq- \ y~ ‘ H, ,4, ..1 ,. - v.41 . a 1,—y'-- . At the last moment, ta; old SAHJlGS of timethy Here obtained from the Albert Dickinson Seed Jomgany. The snugles gave solution resist— ance readings of 8,048 and 8,551 ohms, which ere both considerably below the lover limit of the range covered by the urtiricinlly-mixed samples. This indicated that their perminetion should be expected to be consider— I) F "‘7 duly less than 73.9 per cent. Germinution tests with some of the sonned seeds (200) used for the resistance measurements gave 40 fer cent and 53 her cent germination respectively. It hes been found, by cemparison with germination tests using ordinary dry seeds, that such previously soaked seeds give only aggroximnte results, so that the percentages re— ported are merely indicative of s germination considerably lower than 70 per cent. 1. 5. 6. 8. 9. 10. 11. 12. H .JeDer, O indestiones physioloricue de phaenomenis galvano-msgneticis in corpore humane observatis. Leipsic, 1836. du Bois—Reymond, E. Untersuchunsen uber thierische Electricitat. Berlin, 1849. Rec-like, J. Tetanus - eine ghysioloqische Studie. Leipsic, 1865. Stewart, G. N. Elektrische Leitfahigkeit thierischen Flussigkeiten. Zentrulalutt f. Physiol. 11:332. 189?. 110th, :1er Zentralblatt f. ”hysiol. 11:217. 1897. Bugarssxy, 5t., and Tsnyl, F. Eine Kethode zur Bestimmung des relativ Volume des Glut- Korgerchen uni dos Plasmas. Zentrnlblntt f. Phgsiol. 11:297. 1697. Osterhout, J. J. V. A method of measuring the electrical Conductivity of living tissues. Jour. Biol. Chem. 36:Ho.3, 1918. A comparison of permeability in plant and animal cells. Jour. Gen. Physiol. lzdt9-413. 1919. Antsgonisn between alkaloids and salts in relation to penueability. .Lmrr. Gen. ifirrsiol. 145515-519. 1919. A comparative study on permeability in plants. Jour. Gen. Piysiol. 1:299-304. 1919. Direct and indirect determinations of hermenbility. Jour. Gen. Physiol. 4:375—283. 1923. Green, R. G., and Larson, W. P. Conductivity of bacterial cells. Jour. Inf. Dis. 30:550-558. 1922. 15. 14. 15. 16. 17. 18. 19. Johnson, I. S. 3., end Green, R. G. Conductivity of yeast cells. Jour. Inf. Dis. 64:186-191. 1934. Weshburn, E. V. The measurement of electrolytic conductivity. Jour. Am. Chem. Soc. 58:2451-2460. 1916. An introduction to grinciglcs of physical chexistry. Ch. 3L); 0 1915. Taylor, 7., and Acree, S. F. Studies in the measurement of the electrical conductivity of solutions at different frequencies. Jour. Am. Chem. Soc. 38:2396-2450. 1916. Hibbard, R. P., and Shugmsn, C. T. A simplified aogeratus for measuring the conductivity of electrolytes. mich. Exp. Sta. Tech. Bul. 23. 1915. Green, N. B. The use of the vibration gelvunometer with a BO-cycle alternating current in the measurement of the conductivity of electrolytes. Jour. of Rot. 4:411-416. 1917. Findley, A. Practical Physical Chemistry. 117-180. 1935. ..39— ‘7 '1 'VIII. IXLLYlIVTILflTCJ“ ILLUfVWLITIUnc. Fiwure 2. Scheme of connections of ugpnratus used in these investigations. -4 Fifiure 5. Electrolytic cell of irmersion type used to Heusure solution resistances. Grugh showing th> Solution resistances for I N5 Figure tiuothy of various per cent germination. We czeruve curve was computed for th.se data , W becnxse of the 11$,LI'rC"."I rain-fie over 17.7.31101] the vermin tion BuFCOMtQHUS extend. Fiwure 5. Grsgh shaving the solution resistances for refl ciover of Veriius yer cent germin tion. The curve of best fit Was Computed by the method of "least Squares." Figure 1. Complete setup. 220 6' 60 F S 1 D/(ey -/.ro—- Ac. ’9' r 0 4c. 1 c S g——a Var/able I? Figure .‘3. Diagram of scheme of connections. Figure 3. Electrolytic cell. tances r8515 1011 O Jlut for timothy. LOWlng s« ‘ ooh s Gr Figure 4. 1;; \ \- E 0: 96 fiermhdm 30 £4 57 60 7v 6” 90 Figure 5. Graoh showing solution resistances for red clover. ,..,,..,yr*« , i (.79- '9 "'TITI'ITIH‘JHMEIIMHILilliiflfflflfliiflflflflflmfl“ 3 129