. I ; |y m M m H ‘I II | ‘ x ‘ I t l i \ Ill ( 1 ‘III M w l I I l t ‘ I THS MEASUREMFNT OF DENSiTIES AND DEVELOPMENT OF WORKING CURVES USENG A RECORDING AND A NON-RECORDiNG MICROPHOTOMETER The-sis for the Deg-roe of M Bx MICHIGAN STATE COLLEGE Marjorie jean Lesher 1943 n. I _. 1- "Y This is to certify that the thesis entitled M51 mam w Q\’\A.o~\ Maxob‘g‘r presented by /VUMADVG~JL‘E{DwLA£}Lgr has been accepted towards fulfilment of the requirements for ii 3 degree in )MLO M AM a, Major professor i. ’ . .' "' {rflsm‘ $3,. Inesunmnnnr OF‘DENSITIES AND DEVELOPMENT or WORKING castes USING.A RECORDIDG AND A romancoamm moaornormm ' h: maorie Jeen Leeher L WIS Subtitted to the Graduate School of niohigen State College et Agriculture and Applied Science in per-tie]. fulfilment of the requirements for the degree 91’ mm 025‘ SC IENCI Department of Chemistry ' 1943 a 1,, 3, 1.14.1." L»; l,_i - 51% (i f‘\ (0 ACKNOWLEDGE“! Ihe nriter wishes to express her appreciation to Dr. D. 9. Doing for his guidance and counsel during the course of this investigation; to the Children's Fund of lichigan and to Dr. Icie lacy Hoobler, Director of its Research Laboratory for the fellowship grant which made this Iork' possible; and to m. Jane: E. Brody ef the Children's Fund of Michigan for graphing the plates used on the Leeds and Northrup recording microphetoneter. e‘ 4 C. \D iésox: This investigation was carried out with the eventual analysis of the trace elements in.milk ash in.mind. rho problems considered were primarily photometric and concerned particularly measurements made with a recording and_a non-recording microphotmmeter. SOEUTIONS. The solutions used in this investigation were prepared to resemble,in the major constituents,a solution of an average sample of milk ash and were made on the assumption that the amount of ash used for analysis would correspond to l g. of ash in 5 ml. of solution. Ihe‘base solution contained in one liter: 40.8 g. of potassium added as K01 32.8 g. of calcium added as 03(N03)2.£Hé0 26.0 g. of phosphorus 18.5 g. of sodium } added as nanépod'nzo 2.76 g. of magnesium added as Hg012.6320 27.0_ g. of ammonium chloride‘ 8.0 g. of capper added as CuCl .2H20 .05 g. of bismuth added as Bi(§05)3.5320 In.previous work on solutions of b1010gical ash the addition of ammonium chloride was found to improve arcing conditions.03 For this reason ammonium chloride was used in these solutions, although no tests were made as to whether it had any advantages in this case. Copper was added as an internal standard. the concentration of copper used was determined by preliminary eXperiment and was limited to a relatively high concentraé tion because of the presence of the element in very small amounts in milk ash. Bismuth was also added for an internal standard but was not used. ‘ the solvent for these salts was a solution of equal parts of water and acid, the acid itself being a mixture of equal parts of concentrated nitric acid and concentrated hydrochloric acid. the high concentration of acid used was necessary to keep the salts in solution. All chemicals used were of G.P. grade and the base solution was checked spectrographically to be sure that it contained no aluminum or manganese. a series of solutions varying only in the amount of aluminum.and manganese present was made with the base described above. fhe concentrations used were (in ag./l.): Aluminul - 150, 120, 90, 75, 65, 60, 55, 50, 45, 40, 30, 30. thanganese - 3.75. 5.00, 2.25, 1.88, 1.63. 1.50, 1.58. 1.25. 1.15, 1.00, .75, .50. The range of concentration of aluminum corresponded to its occurrence in milk ash as reported in the literature. the concentration of manganese was somewhat higher than its natural occurrence. INSTRUMENDS. the instruments used for this work were: Dausch and Lamb Littrow Spectrograph no. 4592, Bausch and Lamb Density comparator so. 2108. Eastman Densitemeter model B, and Leeds and Northrup Knorr-llbers ldcrophotometer for Recording Spectrum Line Densities. PREPARATION OF ELECTRODES. national Spectroscopic % in. carbons containing traces of titanium, vanadium, iron, magnesium, and possibly calcium were used for electrodes. tor the lower electrode the rods of graphite were broken into 5 cm. lengths and both ends were used. During the course of the investigation three types of lower electrodes were used: (1) the crater electrode of 6 mm. depth and prearced twenty seconds at 9 amp.; (2) the crater electrode the same as above, but treated with kerosene; and (5) the crater post electrode made with a Dietert cutting tool, with a depth of 7 mm.,and with the center poet 2-2fi-mm. above the rim after prearcing for ten seconds at 9 amp. The upper electrode used with the crater type of lower electrode was pointed with a hand pencil sharpener used especially for this purpose. One carbon was used several times by removing about 5 mm. of the end and reshaping. With the crater post electrode the upper electrode was lI8 in. in diameter, was pointed, and was never used twice. The lower electrode was filled with .1 ml. of solu- tion and was baked for one hour in an even at 145° c. before arcing. Two methods of filling were used--one with the electrode at room temperature, and one with the electrode warm (i.e., filled immediately after removing from a 145° 0. even). When the electrode was filled warm, liquid appeared on the outside of the electrode about 1-2 cm. from the top. However, this liquid dis- appeared and seldom left a residue of salt after baking. It was noted at the beginning of the investigation that, -3- if there was much liquid in the electrode at the time it was put in the even, large amounts of salt collected at points on the inside, outside, and rim of the electrode. For this reason care was taken to evaporate the samples before baking EIGITATION CONDITIONS. A direct current are at 200-290 volts, supplied by a General Electric Generator, was used for excitation. The lower electrode was the anode, the electrode gap was 10 mm., and no lens system was used. The slit width was 40 microns, the shutter was opened before the arc was struck, and it was not found necessary to reduce the light with a rotating sector. The No. 5 opening of the Hartmann diaphragm was used to limit the height of the slit. The exposure time varied with the type of plate and current. Iron was run at 55-60 v. and 5.0 amp., while the solutions were run at 50-40 v. and 9-12 amp., the current dropping and the voltage increasing from this when.most of the sample was burned out and sputtering began. CHOICE OF LINES. The following lines were used in this investigation. Aluminum - 2568.0, 2575.1. 2660.4 A. (Angstrom) Manganese - 2801.1 A. Copper - 2882.9, 5056.1 A. The more sensitive lines of aluminum, 5082.2 A. and 5092.7 A. could not be used because of interferences-- magnesium with 5092.? A. and an impurity in the carbon (probably vanadium) with 5082.3 A. PROCESSING PLATES. Eastman Spectrum Analysis #1 and Eastman #55 plates were used. All processing took place at 18° 0., the developers being Eastman D-ll and D-lS. After developing and fixing, the plates were washed half an hour and then rinsed with distilled water and wiped with a cellulose sponge. EXPERIMENTAL. In the first attempt to make working curves for aluminum and manganese the crater post electrode tilled at room temperature and run at a current of 12 amp. for eighty to one.hundred seconds was used. The results from these trials gave widely scattered points that did not fall along a smooth curve. Repeated efforts using the same excitation conditions failed to show any improvement and when identical samples were run it was found that the logarithm of the relative exposure ratios for the unknown to the internal standard for these samples varied greatly. Hence, it was decided to make a study of excitation condi- tions and types of electrodes in order to improve working curves and the reliability of analyses made from them. Observations were made on the arcing of the different types of electrodes. it was noted that the length of time required to completely burn the sample from a crater or crater post electrode, filled either warm or at room temperature was the same (about 4i-min. at 12 amp.). The sample burned somewhat more rapidly from the kerosene treated electrode. The crater electrode had a very characteristic manner of burning. Per the first twenty seconds or more after arcing wondering of the are on the rim of the lower electrode took place. After this, a period of "steadiness" began, lasting twenty seconds or more. During this time the are seemed to cover the entire rim of the lower electrode and did not wonder. After this period the are again wondered from point to point on the rim of the lower electrode. This ”steady" period depended on the current, beginning later and perhaps lasting a little longer for lower currents. On 9 amp. the I'steady" period began about forty-five seconds after arcing and was somewhat inter- mittent, often lasting throughout the entire ninety second exposure time. The average period of steadiness was about thirty seconds, according to the data from 9 plates representing 80 samples. Iith the crater pest electrode, in the earlier work, the are burned from the post during the first part of the exposure as it theoretically should. With later work, however, it seemed to burn from the rim in the same manner that the crater electrode did. Prom 85 samples, during the first one hundred seconds of burning: 45 did not burn from the post at all. 25 were on the post one to three seconds, 4 were on four to seven seconds, 8 were on eight to ten seconds, 1 was on twenty-five seconds, 1 was on forty seconds, and 5 were on seventy seconds. In all cases where it burned seventy seconds and twenty-five seconds, in some cases where it burned eight to ten seconds on the post,and in most cases of the earlier work not recorded here, salt was piled on the post indicating that this may have been an important factor in the manner of burning. With the kerosene treated electrode the arc wondered rapidly over the rim of the lower electrode and in most cases there was no 'steady' period. When there was a 'steady" period it began approximately twenty-five seconds after arcing and lasted about fifteen seconds. It was shown that after a certain amount of burning the arc began to sputter and with this sputtering, heavy background came in. Thus, from the standpoint of holding back background it would be desirable, if possible, to avoid this sputtering. At 12 amp. with the kerosene treated electrode, sputtering began on an average of seventy seconds after arcing, while with the crater post electrode it began eighty-five seconds after arcing. With the crater electrode filled at room temperature the time of sputtering varied with the current as follows: at 11 amp., after eighty seconds; at 10 amp., after one hundred seconds; and at 9 amp., after one hundred and ten seconds. To find conditions which would give less deviation in the relative exposure ratios for the same sample the excitation conditions listed in Table I were tried. The characteristic curve of each plate was obtained from the step sector pattern of an iron line and plotted as galvanemeter deflection versus logarithm of relative exposure (Log E) as described later in the paper. 'The ‘Log E value for the standard line was subtracted from the Log E value for the unknown line to give the logarithm of the relative exposure ratio, Log(Ex/E,t). The variability of the values of L0g(Ez/E,t) was then calculated for each series from.the equation 2x2 - (2 x12 3 : n - n - 1 S : Variability x a Values of LogiEIIEgt) n a Number of samples In this calculation any value of Log(Ex/Eat) which varied from the mean by an amount greater than 2.6 times the variability was discarded. It was found necessary to omit very few values for this reason. To test to see if the difference in variability from one series to another was significant Snedccor's "I“ test was used at the 5% pointsz) The variability for five pairs of lines is shown in Table II. It was not intended that an exact rating of the different series be made and the data are insufficient for that purpose. However, it would appear that a lowering of the current to 9 amp., using crater -9- .mswa mammosmm m mm.a om m.HH soom smpmho om Hmam mm 2 .mEme mcwmopmm b mm.H om m.HH soom Honcho 00H HwHH.o mo.s .oo.a .oo.H moo.u .mao.u .mmo.- HHo. moo.-.sao.:.moo.u.aoo.: moom so\aoom as omm.-.aam.-.oom.u N.m m.os H.mo .o.oo .o.mm oom.- .omm.u .Nom.s oao. omm.:.mom.:.aam.:.amm.n moow so\ooom as omo..ooo..moo. &N.m Ra.m H.mo .o.oo .o.om «so. .mmo. .omo. who. .oso..oso..sso..mmo. moom so\mamm as d mmfihmw .xss .asg .xos moms .oas .xos zoos .osz .onoo oH ooqom OB Apmm\smvmoq assasn Asmm\xmvmoq amass GOHpmH>mQ & wnfiwqommmhnoo .ocoo mo ownmm umwns> Mom mmSHw> . onaomezmozoo so momma 2H oonaoH>mo sosHmas so ooneoqooqoo ozHoooOZH AsmM\smvooq so MeHonsHm¢> oZHeaoooqso 2H ammo «asp mqmssm HHH mqm¢a electrodes filled warm should give fairly good results for aluminum and manganese, although perhaps not the best results for manganese. It is interesting to note that, as would be expected, the variability of the ratios from two capper lines or from.two aluminum lines is relatively low, but there seems to be little correlation between the variability with two lines from the some element and the variability with two lines from different elements. Thus, while Series A and B have a rather low variability for aluminum to copper or manganese to copper ratios, the variability for copper to copper and aluminum to aluminum is relatively high. Sample sets of data are shown for Series A and B in fable 111. Also in this table are given values for maximum percent deviations in terms of concentration. these were calculated by assuming that the mean value ef’ch{Ex/E.‘) was correct and assigning the true concentration of the solution used to this value. Using the working curve for the pair of lines in question, the concentration corresponding to the extreme values of ch(Ex/Bat) were found and from this the percent deviation from the true concentration was calculated. The sums of the deviations for the data showed a range of8.0% to 15.6% which represents the maximum variation for eight samples. Lccording to Pierce and Nachtrieb a.mean deviation cf 5% in a series of duplicate determinations using graphite electrodes is good and often not better -12- than 10% deviation can be ebtainedfz) Hence, satisfactory results could be expected from the excitation conditions used for Series A and B. fhc conditions finally decided upon were: crater electrede, filled warm, a current of 9 amp., Eastman #33 plates and an exposure time of ninety seconds. this exposure time seemed desirable from the standpoint that the arc was “steady" during much of this period and that no sputtering of the arc occurred. IEABUREMENT 0F LIED BLLGKNESS. lhasurements of line blackness were made with.a Bausch and Lomb density comparator and with a Leeds and Northrup Knorr-Albers microphotometer for recording spectrum line densities. 9n the Banach and Lomb instrument the galvanometer scale covered 38 cm. with increasing scale reading indicating increasing blackness. a total balckness reading was made by shutting off the light from the source with an opaque card. The clear plate reading was made on an unexposed portion of the plate and was kept constant, usually at aero, by adjusting the amount of light striking the photocell. The galvanmmetcr deflection for a line was determined by going back and forth over the line slowly until the maximum reading was obtained. For the calculation of the density of a line ch(l.[I) was used. where the ID measurement was taken as the galvanometer reading for total blackness minus the galvanometer read- ing for the unexposed portion of the plate and the 1 measurement was the galvanometer reading for total -13- blackness minus the galvanometer reading for the line. The Leeds and Northrup microphotcmeter consisted of two units, the Knorr-Albers microphetemeter and the Leeds and Northrop Speedemax Recorder. The plate was scanned on the microphotometer unit and the densities were recorded on logarithmic-scaled paper graduated in densities. The density was defined in this case as Log(do[a) where do was a deflection corresponding to an unexposed portion of the plate and d was a deflection corresponding to an exposed portion. The plate scanning speed used in this work was two millimeters per minute. The two microphotometcrs were found to vary somewhat in the measurement of density. A calibrated Eastman step density tablet was read on the two instruments mentioned above and also on the Eastman densitcmeter, lodel B. In order to use the step density tablet with the Bausch and Lomb and the Leeds and Northrup micro- photometers it was necessary that it be mounted on a spectrographic plate from which the emulsion had been removed. Zero readings, then, were made on the clear glass plate. The results are given in Table IV and are shown graphically in Fig. 1. It will be noted that the values from all three instruments vary, but that within the accuracy of the readings there seems to be a linear relationship between the densities above 0.5. The Bausch and meb values and Leeds and Northrup values are fairly close together, -14- TABLE IV DENSITY OF EASTMAN STEP DENSITY TABLET STRIP #522 AS MEASURED BY THREE INSTRUMENTS Measured Density Step Certified Density Eastman Bausch & Leeds & Lomb Northrup 1 .05 .08 .10 .06 2 .19 .25 .52 .22 5 .54 .40 .57 .45 4 .49 .54 .80 .69 5 .64 .71 1.05 .91 6 .79 .84 1.28 1.15 7 .95 1.00 1.50 1.5 8 1.07 1.12 1.68 1.5 9 1.21 1.29 1.92 10 1.55 1.45 2.12 11 1.50 1.58 2.55 12 1.65 1.72 15 1.80 1.86 14 1.94 2.01 15 2.09 2.15 16 2.24 2.28 17 2.58 2.41 18 2.55 2.57 19 2.68 2.72 ~15- ...q.a J'f i‘clfi.Nerij ,‘Vfld , i—c-vn.o .ttt‘v . H at X23500 stock 25.23% 2.2. mm Qw‘aitmu ct aksoth w . a 2.--- we as. N< 1-3-32 $1110.17}: ---.1‘.--oJoloicllio1J-Il14$¢_¢-Ilf. 1-0: . n . . — m . . , n o... . . . . s . . _ a .. a .5. ...-..... : . . :M _ . .. .. u .. . .. 1 ~ _ . .. 1 . . ., _ -- *v 1.523;...ptzh - - . . U . . o a . V o . .i..fl!.-. .x-.-...... .L .~-—— v - "v n -o..—..- Q6 $6 .tvv-.. . N4 Q6 fi< ¢.~ 9N sue/runs; sN/ mamm was/.4 azmwso sv A1. IS'NJO with the Bausch and Lomb being slightly higher. This condition was consistently found to be the case for step density tablets read on other plates. This did not hold true when lines rather than continuous blackening were read. The density values of lines read on the Leeds and Northrup instrument were equal to or higher than the density from the Bausch and Lomb. Data illustrating this are shown in Table V where lines and a step density tablet from the same plate are used. In comparing the measurement of lines and of continuous blackening mention should be made of the fact that measurement of the step density tablet on the Banach and Lomb instrument was not a scanning process as were all other measurements on the Bausch and Lamb and Leeds and Northrup instruments. Usually several days elapsed between the reading of a plate on the two microphotometcrs. Since this might give rise to misleading results due to a possible change in density upon aging of the emulsion, some plates were reread afterwards on the first instrument used and the tendencies discussed previously were found to hold. a comparison of working curves obtained by using the two micrOphotometcrs was made. it was found that values of Log(E;/Eat) varied slightly between the two instruments. However, the results from one instrument showed no tendency to be high or low, or to give a different slope to the working curve. All variations could probably be accounted for in errors in the original density measurement and in -16- TABLE V DENSITY MEASUREMENTS FROM BAUSCH AND LOMB DENSITY COMPARATOR AND LEEDS AND NORTHRUP MICROPHOTOMETER PLATE NO. 109 Step Density Tablet 41 2575 41 2660 B&L L&N B&L L&N B&L L&N 1.57 1.50 .855 .700 .410 .412 1.55 1.50 .584 ‘ .598 .542 .540 1.52 1.05 .410 .440 .225 .250 .870 .820 .407 .450 .225 .240 .619 .590 .514 .550' .152 .185 .455 .400 .591 .418 .207 .226 .258 .250 .554 .585 .195 .195 .154 .150 .524 .545 .175 .178 .061 .055 .512 .550 .170 .180 .021 .018 .265 .287 .150 .144 .009 .005 .175 .194 .095 .097 .000 .000 -17- TABLE VI VALUES OF LOG(E y/Est) AS OBTAINED FROM BAUSCH AND LOI.lB DENSITY COMPARATOR AND FROM LEEDS AND NORTHRUP MICROPHOTOMETER PLATE NO. 109 Cone. of 41 11 2575/Cu 2885 41 2660/Cu 2885 mg./1. B 5 L L 5 N B 5 L L 5 N 120 .50 .52 .20 .17 90 .41 .44 .12 .11 75 .25 .24 —.02 -.04 85 .22 .25 —.04 -.05 60 .18 .18 —.11 -.08 55 .18 .18 -.09 -.08 50 .15 .14 -.11 —.10 45 .10 .09 —.15 -.15 40 .02 .04 -.21 -.18 50 —.02 —.05 -.27 -.25 20 -.17 —.16 —.56 -.57 Cone. of Mn _Mn 2801/Cu 2885 Mn 2801/Cu 5058 m3 /1 B 5 L L 5 N B 5 L L 5 N 5. 00 .25 .21 .08 .04 2. 25 .11 .10 -.08 —.11 1. 88 .07 .04 -.12 -.12 1. 85 .05 .01 —.17 —.19 1. 50 —.04 —.04 —.19 -.19 L 58 -.02 -.05 ' -.19 —.21 1. 25 —.05 —.05 —.22 -.19. 1.15 -.05 -.05 -.22 -.21 L 00 —.10 -.11 —.50 —.50 .75 —.19 —.20 -.56 —.55 .50 —.25 -.25 —.40 —.58 ~18- and in plotting and making readings from the characteristic curve. Table VI gives samples of values of Log(Ex/Est) cbtained from the Banach and Lomb and from the Leeds and Northrup microphotometers, while Fig. V shows two working curves plotted with values from both instruments. CHARACTERISTIC CURVE. In making the characteristic curve of a plate two problems arose. The first was whether there would be a difference in using the step density tablet or the step sector pattern of iron for this purpose. The step density tablet involved light of the visible range whereas the step sector pattern involved a rotating step sector placed in front of the slit and light of any wave- length recorded by the spectrograph could be used. The results of the study are illustrated by Figt II. It will be noticed that the slope of the curve from the step density tablet is considerably greater than that from the step sector pattern. Thus, the step sector pattern using the same wave length of light as used for analyses is preferable to the step density tablet. The second problem was whether, with the Bausch and Lamb density comparator, the use of Log(Io/I) would be better than the use of galvanometer deflection as the ordinate for the characteristic curve. with ch(I°/I) being a density measurement, the characteristic curve obtained from it would be more like the actual Eurter and Driffield curve. Sets of data were calculated using both types of characteristic curves and the values of Log E as -19- : t a I .-L..-..... e 0 -\ e e t e a - » l . 2 ... . .. 2 . ...-. , - - o .. 2 e s 9 O 45'». Y ass (/I 507) .41/swig F/‘.IZ' L06 or Runr/ v: ExPOOURE (lay. E) obtained from the two curves were compared. Data from Plate No. 87 are shown below in Table VII TABLE VII LOG s VALUES 0132111151) mos DIFFERENT CHLRACTERISTIC CURVES Log E obtained Log E obtained , from curve from curve Galv. R°‘d' F°5(:°/I) using galv. using LogCIOII) deflection From 11 2575 51.7 .857 6.56 6.57 50.5 .756 5.95 5.92 28.9 .655 5.42 5.45 25.9 .520 4.55 4.56 25.0 .486 4.50 4.52 21.0 .562 5.48 5.50 16.0 .245 2.69 2.70 12.1 .171 2.11 2.15 From Al 2660 , 11.5 .157 1.99 2.05 8.6 .114 1.56 1.62 6.0 .076 1.15 1.15 The conclusion that may be drawn is that in the range of densities covered it makes no difference whether Log(I°/I) er galvanometer deflection is used for the ordinate of the characteristic curve. This, of course, assumes that if galvanometer deflections are used the clear plate and total blackness deflections are kept constant while the entire plate is read. Since the method of galvanometer deflections did not involve as_much calculation as the Log(Io/I) method, galvanometer deflections were used for the work where density did not have to be calculated for other reasons. -20- BACKGROUND. A brief study of background was made for the Eastman #55 plate. Background densities were treated the same as line densities, the Log E values being obtained from the characteristic curve in the same manner. The Log R value for the background was subtracted from the Log E value for the line. The difference between the Leg E values for the unknown and standard lines, thus I'corrccted" for back- ground was then taken. When the values of LogQEx/Ea‘) so obtained were plotted against Log Concentration, the points were found to be rather widely scattered instead ef lying along a definite smooth working curve. One explana- tion of this would be the magnification of errors by working on the toe of the characteristic curve. The density of the background ranged from .02 to .14. Since these values fall on a portion of the characteristic curve which has a very small slaps compared with the portion usually worked with, a slight error in background read- ing would give rise to a relatively large error in the Log E value. The background was irregular and its measurement was difficult and relatively inaccurate, especially with.the non-recording microphotometer. Thus, treatment of the small amount of background dealt with in this investigation was not practical with the method used. Samples of working curves made by using the excitation conditions discussed previously and the method of measure- ment and treatment of line blackness discussed in this last section are shown in rigs.iII, IV, and v. The curves -21- in Figs. III and IV are made from measurements on the Bausch and Lamb density comparator, while those in rig. v are made with.measurements from both Bausch and Leah and Leeds and Northrup instruments. . I SUMMARY. Excitation conditions used in the development of working curves were discussed. .Measurement of densities on a recording and a non- recording microphotometer was treated. It was found that actual density values varied somewhat between the two instruments, but the working curves obtained from each were the same. -22- -.—. e .. -- .— o...- 0 v a4if‘ff-7- - - ---~.-.._.-. .- , ' N...- ._-_.._.- ___..._.- F rmms cums MAW - ... : - .i- ; -a---¢—~—- t t o -o a o o u u u u 1.7 to 1.9 2.0- 1.01 Cmmrm (a 1: Jo ”metastases.” :0 cavern TRATION (Milligrams per Lifer) Fig. .HZ’ . \ 0-5 "' A- - w—--———.. 0 ' t e n O o . - d l ’ .... ' a» 5 t 1 ... s . . . , . ~ . . ~ - I I I . . u 1 n . . t . o 1 o . . - . — e -.-~< -—--~ ----»» . -_- --—-- . Q . .- o..—- o...- . . . 7 ° . . - e s . t 0 - O a : ~ . z . .. e . o . . 2. n - e . o . . T . ' I ‘ I 1 ‘ ' 0.4 ' ‘ ‘ 7 v e . ._ - I l .2 o~~ ._.._.- u o . , . I o w I . . 1 . t r _ . s o I D .00... s 00.-.... s- use. H v t - . v , . - .o o - - a . .- .. _-. .._,..-» ......... - o p........-....- - _—.-...... O 0.-....- . -..-. -.. A I . . e .. . 0 ' 1 o I 1 . . g -. . . _ 2 - _ 2.. .-_..... ‘ . . , . . . . . .. . . . . . . s 2 . .e .a ‘-. e -‘ o.-.’