Document 6RnL9n3XqgONVG00BZkYo1yBE

IUin'i rtrJrftrrJj Vol 9. pp. 93? lo 944. Pcrgnnmn Press 1975. Printed in Great Britain. PROBLEMS ASSOCIATED WITH LOW-SOLUBILITY COMPOUNDS IN AQUATIC TOXICITY TESTS: THEORETICAL MODEL AND SOLUBILITY CHARACTERISTICS OF AROCLOR 1254 IN WATER* W. P. Sciiook U.S. Environmental Protection Agency, Gulf Breeze Environmental Research Laboratory, Sabine Island, Gulf Breeze, Florida 32561, U.S.A.t (Received 9 January 1975) Abstract--A theoretical model of the behavior of substances having low water-solubility is presented and discussed with respect to aqueous bioassay. UKraccntrifugal techniques were used in an attempt to study size distributions of Aroclor 1254 aggregates in aqueous emulsions. Results indicate strong adsorption from emulsion by surfaces and a walcr-solubiiily at 20C of less than 0.1 //g 1"1 in distilled waler and approximately 40% of that value in water containing 30g 1 ~1 NaCl. Implications with regard to aqueous bioassny arc discussed. INTRODUCTION the idealized conditions of a two-component sys Laboratory experiments designed to determine the tem--a single solute and a single solvent. A definable effects of chemicals on aquatic organisms require that system should, however, be the starling point of any tiie tests be conducted under conditions which repro-. scientific investigation aimed to arrive at data which duce those present in nature as closely as possible. lead to a quantitative understanding of the behavior In order to accomplish this in a precise and scientific of a compound in water. With these data a more fashion, the physical slate of a compound in an precise attempt can be made to extrapolate from a aqueous dispersion must be known. Convenience, system employed in llic laboratory to the obviously time and other factors have in the past often led to much more complex system present in natural waters. the use of techniques in the laboratory which do not The purpose of this work is to provide a working take into consideration that the solubility characteris theory on the behavior of substances of low water- tics of a compound may possibly affect (lie toxicity solubility and to test this Iheory by investigating the necessitating extrapolation from an apparent toxicity solubility characteristics of Aroclor 1254. established in the laboratory to an expected toxicity under field conditions. In many instances, (he practice of using extrapolation in scientific investigations is THEORY necessary and has proven to be a valuable tool when To explain and predict the characteristics of water- certain conditions cannot be met. However, the range insoluble substances at low concentrations, an through which the extrapolation is carried out must attempt is made here to redefine the basic principles be chosen with great care, because without sufficient underlying a disperse system. No aftempls have been experimental and theoretical justification, a resulting made to include in the definition the somewhat obso extrapolation in this light may well prove to be unrea lete and often vague definitions of emulsions, suspen listic. Since natural water conditions represent a sion, colloids, etc. The characteristics ascribed to each multi-component system, any attempt to understand becoming readily apparent as the theoretical treat it quantitatively must be preceded by a study of the ment of the proposed model continues. system under ideal conditions. While the knowledge In this paper, an ideal or true solution is defined thus gained may or may not be of consequence in as a solute dispersed in a solvent so that any single direct application, it, nevertheless, provides a more molecule of solute is surrounded by enough solvent precise scientific basis for choosing valid limits for molecules to ensure that at any instant all solute extrapolation. molecules are distributed statistically equidistant, The physical State of a compound in water is not assuming a dilution at which interactions between a simple and straightforward phenomenon, even given solute molecules become negligible. The ideal solution, under the conditions described, * Contribution No. 208, Gulf Breeze Environmental Research Laboratory. t Associate Laboratory of lire National Environmental Research Center, Corvallis. Oregon. *> Registered'trademark, Monsanto Company, St. Louis, MO. Mention of trade names docs not constitute endorse ment by the Environmental Protection Agency. is represented by the presence of single solute mole cules. Solute aggregates consisting oT two or more molecules may represent a deviation from the ideal solution because, at least theoretically, these aggre gates could consist of any number of molecules whose behavior would not necessarily coincide with that of 937 , 9.1H W. P. Sciioor u single molecule. For each solute and a single sol vent. there is assumed to exist amongst all aggregates a maximally stable aggregate which, due to its nature, remains statistically equidistant from all other aggre gates for at least a certain period of time. The stability of this aggregate depends solely on the molecularly characterized interactions at the solute-solvent inter phase and on temperature. By definition, a single solute molecule in a disperse system possesses a certain sphere of influence, the nature of which governs (he fate of the solvent mole cules that surround it, which in turn affects the behav ior of tire solute molecule, and thus determines the characteristics of the solute molecule in the system. While precise information is lacking, it is known, nevertheless, that the range orcffecl of a soluic mole cule may extend through several layers of surround ing solvent molecules. This means, of course, an orderly alignment involving either oppositely charged [volar regions or non-polar regions on the solute and the solvent molecules. If this interaction between solute and solvent molecules is of significance, the above defined ideal solution can be visualized, pro vided also that there is no competition among the solvent molecules belonging to respective spheres of influence of two separate solute molecules, The complexity of the situation is increased in cases where tiie interactions between solute and solvent molecules (solute-solvent interactions) become less pronounced, and. as a result, the interactions between solute and soluic molecules (solute-solute interac tions) become more pronounced. This implies that the sphere of influence around the solute molecule is diminished with respect to the solvent molecules which are now no longer attracted to the same degree. As two or more solute molecules start to form aggre gates. the factor oT size of aggregates versus their sta bility in a solvent becomes of utmost importance. A generalized illustration of the size distribution of aggregates that one might expect to find in a susixtiision is shown in Fig. 1. Region "A" describes an area in which the aggregates arc too small to exist independently because interactions in the sphere of influence al that point are such that solute-solute in teractions. which have now become aggregate-aggre gate interactions, arc more pronounced than the aggregate solvent interactions. Therefore, these aggre gates arc expected to coalesce, moving them into region "B'\ which describes a range of aggregate sizes of maximum stability. The aggregate-aggregate inter actions in this range are weaker than in region "A" for that size of aggregate. Region "C' described aggre gates which are loo heavy to remain in suspension for a given period of time and will settle out or break into smaller, more stable aggregates. The exact shape of (his curve and especially that of region "B", depends on how tightly the solvent is held within the Fig. I. Theoretical relative stability of different sizes of aggregates in an emulsion during a given time interval. sphere of influence of the solute aggregate, which is a function of the molecular interactions between solute and solvent. The distribution of different aggregate sizes in terms of molecularly characterized interactions is shown in Fig. 2. The actual equilibrium reaction tak ing place is described in a simplified manner at the top of the figure. The two curves relate the hypotheti cal strength of interactions of solute-solvent (aggre gate-solvent) type and solute-solute (aggregateaggregate) type to aggregate size. The region where the curves cross corresponds to a distribution of aggregate sizes of maximum stability. MODEL Aroclor 1254 was chosen as a model compound because it has been extensively used in bioassay at this laboratory (Duke. Lowe and Wilson. 1970; Nimmo ft a!., 1971a; Nimmo ft a!., 1971b; Hansen cl at., 1971; Lowe el at.. 1972; Walsh, 1972; Cooley, Kellner and Forester, 1972). One approach to estimate quantitatively the solubility of Aroclor 1254 in water and the behavior of its aggregates is to use ultraccntrifugal analysis. This technique permits the selective removal of particles of a certain size. For a spherical particle having a density of (pj and a radius of (r) the molecular weight (mot. wt) is represented by: mol. wt = 4/3rtr5pA'0 (1) where A'0 is Avogadro's Number.* . Two opposing forces if) which determine the fate of a vonndgi* money)* to) 'a* (tOonflfc) ft) (c) --------9 * The equations used are normally found in any text book on physical chemistry, and their reproduction here Fig. 2. Theoretical strength of interaction between solute is intended merely for the convenience of the reader. and solvent. 0SW 025264 e-fi -4'Z STLCOPCB4009219 Theoretical model and solubility characteristics of Aroclor 1254 in water 939 particle in solution: sedimentation / = 4/inr,(p ~ p0)fl and . bouynney / R 6im;, (2) (3) where (p0) is the density oT the solvent, (0) is gravity, and (q) is the viscosity of the solvent. To remove a small particle from an emulsion at a reasonable rate, a force larger than gravity must be applied. Using tbe ullraccnlrifuge, (p) in equation (2) is replaced with (co!x). the angular velocity of the centrifuge rotor (to) times the distance of Iravcl (a) of the emulsified particle.. The rate oT sedimentation during centrifugation is described by: d.x 2rJ(p - poYo'x dr " ` " 90 where (r) is time in seconds to reach equilibrium. Integration yields: 2r'(/i In Xj - In X] (5) The radius of a spherical particle is then given by: ' 9^(ln x2 -- In x i) 1/2 r b -------- ---- -- j -- . 2(p - p0y < . (6) Where (o >= 0.10472 (rev min' ') rotor 0 = g cm" 1 s' 1 p - gcm'3 \ =- cm 1=s Knowing the radius of a particle or assuming a radius, the time necessary to remove the particle from an emulsion is given hy: 9q(ln x2 - In Xj) 2(P - PairV The following arc particle size limits calculated using equation ((>) for given centrifugation times, with q = 8.94 x l()~3 g s" 1 cm" *, V| = 6.7 cm, x2 - 15.3 cm, p ~ p,, >= 0.508gem'5 at 25,000 rev min'1. Time (h) Radius of particle (nra) 1 16.3 2 11,5 3 9,3 4 8.1 6 6.6 ' 8 5.7 The following are particle size limits calculated using equation (6) for given centrifugation times, with 0 = 8.94 x 10'3 gs'1 cm' ', ,\| = 6.00 cm, .\j * 10.73 cm, p - p0 = 0,508 g cm' } at 45,000 rev min"1. Time <h) Radius of particle (nm) 1 7.6 2 5.4 3 4.4 4 3.8 (208,000g mole" '* 636 molecules) 6 3.1 8 2.7 12 2.2 (40,000g mole'1* 124 molecules) * Average mol. w! Aroclor 1254 >= 327 g mole 1 (Hutzingcr, &ife and Zilko, 1972). EXPERIMENTS WITH AROCLOR 1254 Wide-month jars, 30 cm high and 14 cm wide, were used to produce 31. of Aroelot 1254 emulsion per batch. Mechanical considerations concerning the proper physical agitation of Aroclor 1254 and water made it necessary lo use 250 ml of Aroclor 1254 in the jar to submerge the blades of the stirrer. Agitation for 0.5 h at 60'C and 1 MX) rev min"' produced a cloudy emulsion which was allowed to settle for 48 h, when the range of concentration was found to be l-20'mgl'1 and the emulsion became almost clear. This emulsion is referred to as type 1. A second homogenization was carried out by transferring to a jar identical to the one used previously volumes of type 1 emulsion to produce emulsions of 10-300 pg 1' and stir ring I h at 25"C and 1800rcv min' *. This emulsion is referred to as type II. Type 111 emulsions were prepared by taking an appropriate volume of type I emulsion, adding it to a stainless steel blender jar to make a total volume of 500ml, and homogenizing at high speed for 5 min. All centrifugations were performed in a Beckman Model L3 50 ultracentrifugc at 20;'C using SW 50.1 and SW 25.2 rotors. The extraction procedure was that of Schoor (1973). with modifications of the ratio of water to hexane. Evaporation was carried out by placing the hexane extracts in a water bath at 35C and allowing a gentle stream of air lo blow across. This method was found superior to distillation in percentage recovery and lime involved. When the extract volumes had to be reduced to less than 10 ml, dried, pre purified nitrogen was used instead of air. A Hewlett-Packard Model 5700 gas chromatograph with a linear electron-capture detector (f,3Ni) was used for quantitative determination of the Aroclor 1254. The linear ity of this detector eliminated use of different standards at each attenuation or reduction in volume of the sample, both being very lime consuming and subject to errors. An OV-IOl column (2% OV-101 on Gas Chrom Q, 100 120 mesh) was operated at 195C with the detector at 300!'C and the argon-methane (10:1) carrier gas at a Dow rate of 25 ml min'1. Except where noted, measurement was made by comparing lota) peak heights of sample and stan dard. To determine the amount of Aroclor 1254 adsorbed on walls of the 34 ml stainless centrifuge tubes, tbe water phase was decanted and any adhering droplets removed with a disposable pipel. Since acetone injected with the sample was detrimental to the chromatographic column, a sonic probe and hexane were used for removal of Aroclor 1254 from the walls of the tubes. This was necessary because the thin layer of water remaining on the walls shielded the Aroclor 1254 and prevented it from being desorbed into the hexane phase. Sonification emulsified the water at the boundary layer, thus allowing the hexane to contact the adsorbed Aroclor 1254. RESULTS A typical chromatogram of an Aroclor 1254 stan dard in hexane (A) and a hexane extract of a type 1! emulsion (B) is shown in Fig. 3, Some of the II peaks indicated arc multiple peaks. Only peaks J-7 were used to calculate the "total" peak height on which all quanlilations were based. Peaks 8-11 were excluded, because they were often too small to permit accurate calculations. The effect of storage time on Aroclor 1254 emul sions of type 1 and type II is shown in Table 1. There is a fairly rapid initial decrease in Aroclor 1254 in all cases and it appears that a plateau is reached at around 7 This should not he interpreted to 940 W. P. St HOOK (o) Aroclor 1254 Fig. 3. Typical gas chromatograms ($cc text for detailed information). 'mean that solubility is approached at that point, only that perhaps a stable emulsion is reached at that point. The hexane extract of type II emulsion (chromato gram B) indicates a relative reduction in peak height for the early eluting peaks. This phenomenon is better described by the results shown in Table 2. For com parison peak 7 was arbitrarily assigned a relative value of 100%. The results indicate that on standing a type 11 emulsion shows a reduction of the individual peaks, with the early eluting components, or less chlorinated biphenyls (Zitko, 1970). being reduced much more than the late eluting ones. The degree of reduction depends somewhat on the preparation and initial concentration of individual type II emul sions (Table 2). Type III emulsions of comparable "to tal" concentration show a relative distribution of the isomers identical to that of the standard. The distribution of isomers in a hexane extract of the gill tissue of a pink shrimp (Penams duoraruni) exposed to 2.5pgl'1 Aroclor 1254 for 20 days is shown in parentheses at the bottom of Table 2. Because peaks 2, 4 and 7 showed obvious contamina tion, peak 6 was assigned the arbitrary, relative 100% value. The "total" concentration of 3.4 mg kg'1 was based on the total height of peaks 1, 3, 5 and 6, and on the wet weight or gill tissue (blotted to remove adhering water). Filtration of type I emulsion through 450 nm (0.45 fi) Millipore fdters revealed obstructed passage of Aroclor 1254 aggregates smaller than 450nm. Start ing with a I mg I'1 emulsion and changing filters after each filtration, less than 0.01 /igl' 1 of the mater ial remained in the water after 15 passages. Since aggregates in the starting emulsion were most likely smaller than 450 nm (calculations using equation (1) Table !. Effect of storage time on amount of Aroclor 1254 remaining in the water phase IfSS-Tdeya)___________ uft/l Aroclor 1254 Type I Type II 0 2 5 ^6 S 9 13 IS 19 20 21 23 26 26 33 34 41 43 2300 502 463 301 115 113 112 97 67 78 426 355 350 280 50.2 286 23.6 11.3 123 98.5 54.7 48.1 44.5 6.7 7.1 6.5 7.7 7.4 15.5 6.8 Table 2. Isomer distribution of Aroclor 1254 type 11 emulsion after standing for various periods of time in 3 1. glass bottle X Peak Height I------------------ Tie Total cone Peak Nutabers 2 (day*) (M&/1)1 1 2 3 4 5 6 7 2 286 76 93 95 95 98 104 100 9 123 79 78 89 94 98 99 100 13 90.5 79 79 S3 98 99 96 100 19 54.7 72 75 85 93 91 95 100 20 58.1 64 70 00 89 92 94 100 21 44.5 61 65 82 90 99 93 100 61 15.5 37 41 56 70 77 88 100 21 13.4 16 27 44 55 76 87 100 33 3.6 12 21 39 45 64 82 100 36 1.6 9 10 31 46 63 100 (3.A pp) (u) (80) (87) (100) ^Calculation. are baaed on the relative height of peek 7 (eet belov). 2Peek nuabara are ehovn on the chronatograa In Fig. 1, DSW 025266 STLCOPCB4009221 Theoretical model mid solubility characteristics of Aroclor 1254 in water 941 lead to roughly 10' times the average molecular weight of Aroclor 1254), the Aroclor 1254 must have been adsorbed on the filler. This was also evidenced by the fact that the filter paper turned slightly trans parent after the first passage during which about 95% of the material was removed from the emulsion. The first centrifugation experiments were carried out by centrifuging 180ml of 42/tgl"*1 Aroclor 1254 type II emulsion in 60 ml polyncetatc centrifuge tubes for 60 min at 107,000 x y (maximum). At an 85% total recovery the following distribution wtis found: acetone extract of tubes hexane rinses of tubes top 50ml water phase bottom 10 ml water phase 66% 18% 5% 11% . The low recovery (85%) was probably due to incom plete extraction of the tubes in spite of refluxing with acetone. Polyallomcr1 centrifuge tubes were tried next. When 180ml of 286pgr1 type 11 emulsion were cen trifuged in 60 ml Polyallomcr lubes for 60 min at 107,(XX) x g (maximum) the following distribution was found: acetone extract of lubes hexane rinses of tubes top 25 ml water phase bottom 35ml water phase -- 22% 0.5% 0.6%. These percentages were based on the total amount of starting material, i.c. assuming 100% recovery in stead of the 85% in the ease of the polyncetatc tubes. Extraction of the Polyallomcr tubes by refluxing with acetone produced too many interfering peaks on the chromatogram, making complete recovery calcula tions impossible. Direct adsorption on Polyallomcr Table 3, Adsorplion of Aroclor 1254 type II emul sion on Polyallomcr centrifuge lubes on standing Aroclor 1254 (pg/*) Time (hr.)_______ __________________ In voter phage 0 125 1 86 72 3.3 0 45 1 35 3 27 tubes was achieved by permitting type II emulsions to sit undisturbed in the tubes. Table 3 shows the outcome for two differ cut concentrations. To permit recovery and study of the material adsorbed on surfaces, 34ml stainless slccl centrifuge lubes were used Tor static tests, as well as for ullracenIrirugal analysis. Table 4 shows the amounts of Aroc lor 1254 adsorbed on the wall or a stainless steel cen trifuge tube in relation to starling concentration and time. The amounts adsorbed from the 14 and 2pgr` emulsions were greater than that adsorbed from the I13pgl"' emulsion during the same time period. It should be pointed out that 0.100pg of Aroclor 1254 adsorbed as a monomolccular layer per tube repre sents about 2% of the minimum area available. The calculated inside area of a stainless steel centrifuge tube was 60.8 cm2. 3This area must be considered minimum because the surface was assumed to be ideally smooth, which certainly is not the case. How ever, for the approximations involved, this figure was used. A simple calculation using equation (I) yields 0.613 nm2 for the cross-sectional surface area of an average Aroclor 1254 molecule using the average molecular weight of 327 (Hutzinger et at., 1972), and p = 1.505gcm"J (W. B. Papageorge, Monsanto Company, St. Louis, Missouri, persona! communica tion). Utilizing a molecular model with tire phenyl groups at right-angles to each other and bond length (Pauling, 1940) as the basis for calculations, a cross sectional area of 0.643 tinr for the fully chlorinated and 0.356 nm2 for the unchlorinated or biphenyl molecule was obtained, Values falling between are not linearly related to amount of chlorination. Using 0.613 nm2 as an approximate, average cross-sectional area, 0. lOOpg of Aroclor 1254 occupies 1.13 cm2 in the form of a mononrolecular layer. This corresponds to approximately 3pgl_1 in a 34 ml stainless steel centrifuge tube. It can be seen that even at 50% adsorption from a 3pgr1 emulsion only about 1% (maximum) of the available surface area is occupied, and surface saturation was not a factor. The amounts of Aroclor 1254 in the form of emul sions of type II and type III adsorbed on the walls of the stainless steel centrifuge tubes are shown in Table 5. There is a difference in adsorplion of the two different types of emulsion in the absence of NaCl. At least for type III emulsions, the introduction i Table 4. Adsorplion of Aroclor 1254 on stainless steel centrifuge lubes as a function of lime and concentration Time (tiro) 0.5 } 1 16 1 7 Tolul Aroclor 1254 type 11 emulsion Wflt-er S.S. tubTT- X adaovbed (ifg) (ifg/i) (liS)......... (uk/jL). <,,X) 3.83 3.83 3.83 3.83 0.48 0.06 113 m 113 m 14 2 3.63 3.31 3.20 3.14 0.35 0.03 107 97 94 92 10 1 0.18 0.30 0.33 0.51 0.08 0.02 5 9 13 16 23 67 bstatnleiis steel centrifuge tubes. DSW 025267 STLCOPCB4009222 942 W. P. Sc hook Table 5. Adsorption of Aroclor 1254 on stainless steel centrifuge lubes tlM (hr*) d/r Arocl or 12541 adsorbed Type 21 Enulaiofi Type III fruition 0 g/l NaCl 30 r/1 NaCl 0 r/1 NaCl 0.5 1.0 2.0 4.0 19 22 0.19 0. JO 0.33 0.42 0.09 0.10 0.14 0.19 0.39 o.io 0.14 0.41 tData adjuatad to 4.00 eg total atartlng aaiount. Table 6. Centrifugation of Aroclor 1254 in water of varying salinities at 69.000 x g (maximum) Tiw (hr*) Mg/l Aroclor 1254 reaftinina in water ph**eJ g/l NaCl 0 15 JO 0.5 X 3.9 7.1 6.0 1.0 12.5 6.6 4.9 2.0 7.2 4.6 2.9 ^Started with 50 eg/l Type III maulalon, of 30gI"1 NaO appears to have no effect on the amount of Aroclor 1254 adsorbed. However, centrifu gation reveals a difference in the size of the aggregates formed in the presence of NaO, as shown in Table 6. In comparison with an Aroclor 1254 standard, the relative distribution of the isomers in emulsions of type 11 and III is quite different, as shown in Tables 7 and X. However, in all cases the adsorbed Aroclor 1254 had a higher percentage or early cluling (gas chromatography) isomers than did that which remained in solulion, DISCUSSION Tiic original intent for conducting the work de scribed was to find the absolute solubility of Aroclor 1254 in fresh and salt water. This, unfortunately, was not completely accomplished to any accurate degree, because a scries of significant problems occurred at the beginning of the centrifugation experiments. Re covery of Aroclor 1254 after centrifugation was low' and, hence, led to the discovery that adsorption occurred on the walls of the polyacctate centrifuge tubes as well as on Polyallomer and stainless steel centrifuge tubes. Ultimately, only the stainless steel centrifuge lubes were used in the adsorption and ultraccntrifuga) studies. The apparent disappearance of early eluting isomers, such as shown in Table 2, has been observed by others. It was found to occur in the eggs of tiie double-crested cormorant and regarded as possibly due to metabolic breakdown (Hulzingcr et al.. 1972). Similar behavior in the carcasses of bobwhite quail after exposure to Aroclor 1254 was observed and believed to be because of isomeric transformations (Bagley and Cromartie, 1973). Application of Aroclor 1254 to different types of soil showed a reduced recov ery of the early eluting, lower chlorinated biphenyls (Iwata, Westlake and Gunther, 1973), and it was pos- Table 7, Distribution of isomers of Aroclor 1254 type U emulsion on stand ing in stainless steel centrifuge lubes Storage (d*y) Hr* in tube tm/i t Peak helRhta* Peak nunbet* 1 Z 34 5 6 7 1 0 310 vater 93 90 90 99 98 100 iqo 5 0 22S water 53 71 73 91 98 96 100 2 97 utter phase 49 67 69 63 100 100 100 2 12 adeorbed 96 106 103 127 119 100 1O0 e 0 112 vater 51 67 71 82 96 97 100 2 102 water phase 48 66 68 79 98 98 100 2 6.0 adsorbed 69 82 85 104 107 100 100 13 0 97 water 47 64 68 81 97 98 100 2 86 water pheae 43 59 66 78 92 96 100 2 6.1 adsorbad 47 69 77 94 101 96 100 tcoafarad to atandard Aroclox 1254 (Fig. 1). Calculations *r* baaed on the relative height* of peak 7. . 2P**k nunbara are ahovn on tha chromatograa In Pig. 1. OSW 025268 STLCOPCB4009223 Tlicorciical model and solubility cliariicierisiics of Aroclor 1254 in water Table 8. Distribution of isomers in tbc adsorbed fraction of Aroclor 1254 type 111 emulsion on standing in stainless steel centrifuge tubes NeCl hre In water phase adsorbed jCb/a)__ tube (i'k/D (fa) 1 % Peak heifthte* PPepaslki nnuiimmtbiperr^`d 2 34 5 6 02 47. 0.122 149 127 135 130 98 100 30 X 46,9 0.075 144 121 129 129 105 100 0 22 39.7 0.190 139 118 1X3 122 127 100 ^Compared to atandard Aroclor 1254 (Pig. 1). Calculations are based on tile relative heights of peak 6. 2PeaV numbers are sltovn on the chromatogram In Fig. 1. 941 litliited that this may have been due to evaporation from the soil, My studies did not substantiate the observations by Zitko (1970) that when Aroclor 1254 emulsions arc centrifuged the dissolved fraction is ridler in the lower chlorinated biphenyls than is the original preparation. However, the difference could be dtie to the method of the preparation of his emul sion, which was similar lo my type 111 emulsion. In both type II and type 111 emulsions the distribution of isomers in the Witter phase shows a loss of the lower chlorinated biphenyls on standing (Tables 7 and 8), This loss was accounted for in all eases by adsorption on the stainless slcel centrifuge tubes, the "losl" lower chlorinated biphenyls always being found in the adsorbed fraction. Thus, at least from water emulsions of Aroclor 1254, loss of the lower chlor inated biphenyls is due to their relatively greater affinity for surfaces. The published values for solubility of Aroclor 1254 in fresh and salt water of 2-3 mg I-1 and 1 1.5 mg I'1, respectively (Zitko, 1970), appear much too high. A conservatively high estimate based on my ultraccntrifugal experiments indicates the average solubility of the isomers to be less than 0.1 pg 1"' for fresh wnler and approximately 0.04 /rg l'1 (calcu lated from Table 6) in water containing 30g 1" 1 NaCJ. It is extremely difficult, in my opinion, to obtain an tibsolute value for the true solubility of the average molecular weight isomer of Aroclor 1254. The prob lem lies in the fact that at low concentrations, long centrifugation times (in excess of 12 h at 243,000 x g (maximum)) theoretically arc necessary to eliminate aggregates from the emulsion. At the low con centrations necessary to eliminate undesirable stirring back after completion of the centrifugation (Bowman, Acrec. and Corbett, I960), adsorption on the walls of the stainless steel centrifuge tubes (67% at 2mgl_1 for 2 h. Table 4) makes it all but impossible to employ ultraccnlrifugation for extended periods of time. It appears that at least in the ease of type III emul sions the adsorption from water emulsions containing Oand 30gT 1 Nad was the same (Table 5), although the rate of sedimentation was quite different. The explanation for this lies in the fact that the size of the Aroclor 1254 aggregate is much larger in the pres ence of salt and, while this is not apparent at ) x g, the larger aggregates are removed more quickly from the salt-containing emulsion during ullraccntrifugalion, This agrees very well with my hypothesis that a larger aggregate is more stable under the given con ditions and in the presence of salt, which is conducive lo greater solute- solute (aggregate-aggregate) interac tion. SUMMARY An extrapolation from the theory presented sug gests that the use of "carriers'' be continued with cau tion, because of two independent effects dial may be present. One effect can most simply be described as an alteration of the aggregate-solvent interactions by "carriers" forming transition-like links between aggre gates and solvent molecules. In such a fashion, solute aggregates are surrounded by "carrier" molecules, thus enhancing the ability of the aggregate to remain in a slable emulsion by permitting greater solute- sol vent interaction. This can be illustrated graphically in Fig. I by enlarging region "B" over a greater range of aggregate sizes since some aggregates previously belonging to regions "A" and "C" now become more stabilized. It may also be visualized by flattening the two curves in Fig. 2, thereby extending their- region of overlap. Thus, when added with a "carrier", more of an insoluble compound may be introduced into a stable water emulsion. The other effect nitty be due to possible interference with the uptake of a test com pound by an organism. Any such uptake must by necessity be preceded by an adsorption to a surface of (he organism such as the gills in a fish. If at this time the "carrier" molecules, which arc located at (he surface of the aggregate, affect the actual process of adsorption in any way, there will be a resultant change in (he rale of transfer of the compound into the organism. If the toxicity is related to rale of uptake, there will be a concomitant change in toxicity. This study shows, both theoretically and experi mentally, that in so far as physical interactions tire concerned, emulsions differing in degree of dispersion and stability can be formed, depending on the method or preparation and subsequent treatment. Conse quently, the following questions should lie answered before conducting bioassays in disperse aqueous sys tems. (a) What tire the solubility characteristics of the compound under investigation'.* (b) To what extent arc these characteristics related to field conditions'.' DSW 025269 STLCOPC 941 W. P. SCIIOOR (c) Mow ctm (lie solubility charactcrislics and Held conditions be best simulated in the laboratory? Such information would undoubtedly result in more precise and relevant data on acute toxicity as well as long-term cfTccls regarding aqueous bioassay of water-insoluble lest compounds. Acknowledgements The author thanks Messrs. D. Lamb and W. Burgess for assistance with the analytical work. Mr. A. J. Wilson. Jr. for the chromatographic column packing. Dr. D. R. Nimmo for the shrimp exposed to Aroelor 1254. and Dr. Ralph Birdwhistcll, Dean, the School or Chemistry. University of West Florida, for reviewing the manuscript. REFERENCES Bagiev G. L. and Cromartic E. (1973) Elimination pattern of Aroelor 1254 components in the bobwhite. J. Chro ma. 75. 219-226. bowman M. C.. Acrcc F. Jr. and Corbett M. K. (I960) Solubility of carbon-14 DDT in water. J. agric. Fd Chem. 8. 5, 406 408. ' Cooley N. R., Kellner J. M. Jr. and Forester J. (1972) Mirex and Aroelor 1254: Effect on and accumulation by Telrahymena pyriformis Strain W. J. Protozool. 19, 4.' 636 638. Duke T. W,, Lowe J. 1. and Wilson A. J. Jr. (1970) A polychlorinated biphenyl (Aroelor 1254) in the water, sediment, and biota of Escambia Bay, Florida. Bull. Environ. Contain Toxicol. 5. 2. 171-180. Hansen D. J,, Parrish P. R.. Lowe J. I.. Wilson A. J. Jr and Wilson P, D. (19711 Chronic toxicity, uptake and retention of Aroelor 1254 in two estuarine fishes. Bull. Environ. Conitnn. Toxicol. 6. 2, 113 119. Hulzingcr O., SaTc S. and Zilko V. (1972) Polychlorinated biphenyls. Analahs Res. Notes 12. 2, I II. Iwatn Y., Westlake W. E. and Gunther F. A. (1973) Vary ing persistence of polychlorinated biphenyls in six Cali fornia soils under laboratory conditions. Bidl. Environ. Contain. Toxicol. 9. 4, 204 211. Lowe J. 1.. Parrish P. R.. Patrick J. M. Jr. and Forester J. (1972) F.ITccls of the polychlorinated biphenyl Aroclor 1254 on the Aniorican oyster (Crassoslrca virginica). Mar. Biol. 17. 3. 209-214. Nimmo D. R., Blackman R. R., Wilson A. J. Jr. and Fores ter J. (1971a) Toxicity and distribution of Aroelor 1254 in the pink shrimp (Penaeus duorarum). Mar. Biol. 11, 3. 191-197. Nimmo D. R,, Wilson P. D.. Blackman R. R. and Wilson A. J. Jr. (1971b) Polychlorinated biphenyl absorbed from sediments by fiddler crabs and pink shrimp. Nature, Land. 231, 50-52. Pauling L. (1940) Nature of the Chemical Band, p, 164. Cornell University Press, Ithaca. Schoor W. P. (1973) In vim binding of p.p'-DDE to human scrum proteins. Bull. Environ. Contain. Toxicol. 9, 2, 70 74. Walsh G. E. (1972) Insecticides, herbicides and polychlor inated biphenyls in estuaries. J. Vl'nsfi. Acad. Sci. 62, 2, 122- 139. Zitko V. (1970) Polychlorinated biphenyls solubilized in water by nonionic surfactants for studies of toxicity to aquatic animals. Bull. Environ. Contain. Toxicol. 5, 3. 219-226. xI II I1I t DSW 025270 STLCOPCB4009225