OFFICIAL PUBLICATION OF THE VIRGINIA ACADEMY OF SCIENCE

THE VIRGINIA JOURNAL OF SCIENCE

EDITOR

Werner Wieland Biological Sciences Department University of Mary Washington Fredericksburg, VA 22401 Phone: (540)654-1426

James H. Martin Dept, of Biology - PRC J. S. Reynolds Comm. Col.

BUSINESS MANAGER

P.O. Box 85622 Richmond, VA 23285-5622 Phone: (804)523-5593

©Copyright, 2005 by the Virginia Academy of Science. The Virginia Journal of Science (ISSN:0042-658X) is published four times a year (Spring, Summer, Fall, Winter) by the Virginia Academy of Science, 2500 W. Broad Street, Richmond, Virginia 23220-2054. The pages are electronically mastered in the Parham Road Campus Biology Department of J. Sargeant Reynolds Community College. The Virginia Academy of Science and the Editors of the Virginia Journal of Science assume no responsibility for statements or opinions advanced by contributors.

Subscription rates: $35.00 per year, domestic and foreign. All foreign remittances must be made in U.S. dollars. Most back issues are available. Prices vary from $5.00 to $25.00 per issue postpaid. Contact the Business Manager for the price of a specific issue.

Changes of address, including both old and new zip codes, should be sent promptly to the following address: R. Gerald Bass, Executive Officer, Virginia Academy of Science, 2500 W. Broad Street, Richmond, Virginia 23220-2054. All correspondence relating to remittance, subscriptions, missing issues and other business affairs should be addressed to the Business Manager.

For instructions to authors, see inside of back»cover.

VIRGINIA JOURNAL OF SCIENCE

OFFICIAL PUBLICATION OF THE VIRGINIA ACADEMY OF SCIENCE Vol. 56 No. 3 Fall 2005

TABLE OF CONTENTS

ARTICLES PAGE

Prospects of Kenaf as an Alternative Field Crop in Virginia. 1 15

Harbans L. Bhardwaj, and Charles L. Webber III.

Fish Consumption Patterns of Populations in Vicinities of Lake Kastoria and Lake Pamvotis, Greece. Eugene G. Maurakis,

David V. Grimes, and Dimitra Bobori. 121

Assessment of Human Health Risks from Chemically Contaminated Lake Fishes In Greece. Eugene G. Maurakis, David V. Grimes,

Dimitra Bobori, Rob Hale, and Jennifer Jones. 141

EDUCATION SECTION ABSTRACTS

155

APR J 6 2111)6 ^

OeRARltb

f 'V r

Virginia Journal of Science Volume 56 Number 3 Fall 2005

Prospects of Kenaf as an Alternative Field Crop

in Virginia'

Harbans L. Bhardwaj^ Agricultural Research Station, Virginia State University, Petersburg, VA 23806 and Charles L. Webber III, USDA-ARS, Lane, OK 74555.

ABSTRACT

Kenaf {Hibiscus cannabinus L.), a warm-season annual plant, has shown potential as an alternate source of fiber in the United States. Although preliminary research has indicated feasibility of kenaf production in Virginia, production details are lacking. Field experiments were conducted during 1 995 and 1996 to determine optimal row spacing and fertilizer needs, and to compare available kenaf cultivars. Although results indicated that differences in dry matter yields from four row spacings (30, 60, 90, and 120 cm) and four rates each of N, P, and K fertilizers (50, 100, 150, and 200 kg-ha'^) were not statistically different, the yields were adequate ranging from 8.8 to 16.0 t-ha'' with an average yield of 12.5 t ha k Dry matter yields for narrow-leaf cultivars proved superior to broad-leaf, and the overall results demonstrate that kenaf can be easily produced in Virginia.

INTRODUCTION

Kenaf {Hibiscus cannabinus L.), a relative of cotton {Gossypium hirsutum L.) and okra {Abelmoschus esculentus L.), is a warm-season annual plant that originated in northern Africa and has been used as a cordage crop for many years in India, Russia, and China (Dempsey, 1975). Kenaf research in the USA began during World War II to supply cordage material for the war effort (Wilson et al, 1965). During the 1950s and early 1960s, it was determined that kenaf was an excellent cellulose fiber source for a large range of paper products (newsprint, bond paper, corrugated liner board, etc.). It was also determined that pulping kenaf required less energy and chemical inputs for processing than standard wood sources (Nelson et ah, 1 962). More recent research and development work indicates that kenaf is also suitable for use in building materials (particle boards of various densities, thicknesses, with fire and insect resistance), absorbents, textiles, livestock feed, and fibers in new and recycled plastics (Webber and Bledsoe, 1993).

These observations indicate that kenaf could be potentially grown in Virginia to diversify cropping systems, to provide alternative materials for paper mills, and to meet varied industrial needs. Virginia State University’s New Crops Program, established in 1991, initiated a kenaf research project in 1992. The objectives of this project were

Contribution of Virginia State University, Agricultural Research Station Journal Article Series No. 247. The use of trade names or vendors does not imply approval to the exclusion of other products or vendors that may also be suitable.

^ Corresponding Author, E-mail: hbhardwj(@vsu.edu

116

VIRGINIA JOURNAL OF SCIENCE

to conduct preliminary production research and to determine the feasibility of kenaf production in Virginia. Research conducted in Virginia during 1992-1994 indieated that kenaf has signifieant potential as an alternate erop in Virginia (Bhardwaj and Webber, 1994; Bhardwaj et ah, 1995). However, information regarding desirable agronomic practices such as cultivar selection, fertility requirements, and plant densities, specifieally for Virginia was not available. Therefore, experiments were eonducted to identify: (1) high yielding varieties, (2) optimum levels of nitrogen, phosphorous, and potassium fertilizers, and (3) ideal row spaeing.

MATERIALS AND METHODS

Three experiments were eonducted during each of 1 995 and 1 996 at Randolph farm of Virginia State University, loeated in Ettriek, Virginia (37® 14' N Latitude and 77® 26 W Longitude) at an approximate elevation of 71 m. The soil type was an Abel sandy loam (fine loamy mixed thermic Aquatic Hapludult) soil that typically has a pH of 6.1 to 6.4.

In the first experiment, four inter-row spaeings (30, 60, 90, and 120 cm) were evaluated with two kenaf eultivars: “Everglades 41" (A kenaf variety with broad leaves) and “Everglades-7 1 " (A kenaf variety with narrow leaves). Three replications of a split-plot design with varieties in main plots and row spaeings in sub-plots were planted on May 22, 1995, and May 20, 1996. Each plot consisted of three rows with a 60 cm spacing between sub-plots. These plots received 100 kg ha’’ each of nitrogen (N), phosphorous (P), and potassium (K). In the second experiment, four rates (50, 100, 150, and 200 kg ha ' ) each of N, P, and K, were evaluated with Everglades 41 variety in four replications of a split-plot design with N in main plots, P in sub-plots, and K in sub-sub-plots. Each plot consisted of three rows with inter-row spacing of 75 cm with one row left blank between the plots. These experiments were planted on May 23, 1995, and May 20, 1996. In the third experiment, 21 kenaf eultivars were planted on May 23, 1 995, and May 21,1 996, in a randomized eomplete bloek design with three replieations. Each plot consisted of three rows with inter-row spaeing of 75 em. These plots reeeived 100 kg-ha'’ eaeh of nitrogen (N), phosphorous (P), and potassium (K).

Approximately 100 seeds of each cultivar were planted in eaeh 3 m long row. In eaeh experiment, weeds were controlled with a pre-plant-incorporated application of 1.5 1- ha'' of trifluralin herbicide. These experiments were not irrigated. Data were recorded for dry matter yield and plant height from samples harvested manually at the ground level after a hard freeze in early January had effectively killed the plants. During 1995, a 1-m sample was taken from the middle row of each plot in each experiment; and in 1996, a 2-m sample was harvested. After a two-month storage period, meant to stabilize the moisture content to a eonstant value and to dry the material, the harvested material was measured and the yield calculated in t-ha''. All data were analyzed using General Linear Models procedure of SAS (SAS, 1996).

RESULTS AND DISCUSSION

Row-Spacing: The differenees in dry matter yield, averaged aeross two eultivars, for the four row spaeings were not significant (Table 1 ). However, the closer spacing of 30 cm between rows showed a numerically higher yield of 1 1.1 t-ha''. The dry matter yields of Everglades 41 (8.2 t-ha'') and Everglades 71 (8.6 t-ha'') were also

KENAF PRODUCTION IN VIRGINIA

II7

TABLE 1 . Effect of row-spacing on kenaf dry matter yield and plant height during 1 995 and 1 996 at Ettrick, Virginia.

Row Spacing

Dry Matter (t ha ')

Plant height (em)

30 cm

11.1 a*

233.1 a*

60 cm

7.2 a

233.4 a

90 cm

7.2 a

231.3 a

120 cm

8.0 a

243.7 a

Mean

8.4

235.4

* Means across two cultivars (Everglades 41 and Everglades 71) and three replications each during two years. Means followed by similar letters are not different according to Least Significant Difference (5% level).

statistically similar (data not shown). The interactions between row spacing and cultivars for dry matter yield and plant height were not significant. The row spaeing effeets on plant height were also not significant. This research demonstrates kenaf s adaptability to varying plant densities. However, averaged across all row spacings, plants of Everglades 71 kenaf variety were taller (242 cm) than those of Everglades 41 kenaf variety (229 cm). No data were reeorded on stalk diameter, however visual observations indicated that stalk diameter in eloser row spacings was less than that of widely-spaced rows. Since a possibility of using kenaf as a forage crop exists, the closer row spacing may be desirable as it would reduce the woody component of kenaf harvested at green stage for feeding the livestock. However, the economics of kenaf seed would need to be considered. Kenaf, being a tropical plant, does not produce seed in the United States. Most kenaf seed is produeed in Mexico or Carribean locations. Use of eloser row spaeing would entail more seed and would inerease production costs.

Nutrient Needs: The dry matter yields and plant heights, following applieation of four rates (50, 100, 1 50, and 200 kg ha'^) eaeh of N, P, and K, are presented in Table 2. The kenaf dry matter yield and plant height did not differ significantly in response to fertilizer rates. However, the highest dry matter yield of 1 1.4 t-ha'^ was obtained upon application of 50 kg-ha’’ N. The residual N content in the experimental area was approximately 14 kg-ha ', therefore, kenaf needs up to 64 kg-ha"' N for optimal production. The response of kenaf to P and K applications was not significant. Previous observations (Rangappa et ah, 2002) have indicated that soil at this experimental site, whieh is generally considered to be typieal of most soils in the Southern Piedmont region in Virginia, contains approximately 54 to 77 mg kg ' P and 52 to 64 mg-kg’’ K. These levels, generally, provide adequate P and K for most erops, and positive responses to additional applieations of these nutrients are not very common. These results indicate that the nutrient needs of kenaf are quite modest.

Varietal ’Evaluations: The mean dry matter yield from 2 1 kenaf varieties was 12.5 t ha'^ (Table 3) which compares favorably with kenaf yields reported from other areas in the United States. The dry matter yields ranged from 8.8 to 16.0 t ha ', respectively for GR 2563 and 78-18RS-10 kenaf varieties. The plant height varied from 229.7 to 288.6 cm, respectively for Tainung #1 and 78-18-GS-3 kenaf varieties. A signifieant positive correlation (+0.28, P=0.001) indicated that taller plants resulted in higher dry matter yields.

118

VIRGINIA JOURNAL OF SCIENCE

TABLE 2. Dry matter yield and plant height of kenaf following four rates eaeh of N, P, and K fertilizers during 1995 and 1996 at Ettrick, Virginia.

Fertilizer

Dry Matter Yield (t ha ')

Plant Height (cm)

Rate

N

P

K

N

P

K

50 kg ha''

11.4 a*

10.6 a

10.8 a

277.6 a

274.7 a

270.7 a

100 kg ha'

10.9 a

11.3 a

11.8a

273.8 a

277.2 a

282.6 a

150 kg ha''

10.6 a

10.6 a

10.3 a

277.2 a

276.2 a

277.9 a

200 kg ha '

10.6 a

11.0a

10.5 a

278.8 a

279.1 a

276.2 a

Mean

10.9

10.9

10.9

276.8

276.8

276.8

*Means followed by similar letters are not different aceording to Least Significant Difference (5% level). The interactions between N, P, and K were non-significant. The means of individual nutrients were obtained from averaging over all rates of other two nutrients i.e. the mean of N is averaged over all rates of P and K, the mean of P is averaged over all rates of N and K, and the mean of K is averaged over all rates of N and P.

TABLE 3. Dry matter yield and plant height of 21 kenaf cultivars when grown during 1995 and 1996 at Ettrick, Virginia.

Variety

Dry Matter Yield(t-ha'')

Plant Height (cm)

Leaf Shape

78-18RS-10*

16.0**

270.0*

NaiTow

Everglades 71

14.3

266.3

Narrow

45-9

14.3

257.7

Narrow

SF 192

14.3

265.2

Narrow

15-2

14.3

278.5

Narrow

KK 60

14.1

280.3

Nan'ow

78-18GS-3

13.7

288.6

Narrow

Gautemala 5 1

13.5

267.0

Narrow

SF 459

13.1

259.2

Narrow

Tainung #1

13.0

279.3

NaiTow

Gautemala 45

12.4

229.7

Broad

C 2032

12.4

247.2

Broad

Everglades 41

12.0

239.2

Broad

7N

12.0

253.8

Broad

Tainung #2

11.9

284.3

Narrow

Guatemala 4

11.7

258.0

Broad

C-108

10.9

245.8

Broad

Indian

10.6

229.8

Narrow

Cubano

10.5

274.8

Broad

Guatemala 48

9.1

270.0

Broad

GR2563

8.8

241.8

Broad

Mean

12.5

261.3

-

LSD(.05)

3.4

31.2

-

* These are the names of kenaf varieties that are assigned by developers of these varieties.

** Means across two years and three replications per year. The year x cultivar interaction was non¬

significant.

KENAF PRODUCTION IN VIRGINIA

119

Among the 21 kenaf varieties evaluated, 12 had narrow leaves and 9 had broad leaves (Table 3). A comparison of narrow-leaf shape group of kenaf varieties with broad-leaf shape group of kenaf varieties indicated that narrow-leaf group had a significantly (5% probability) higher yield and significantly taller plants (13.6 t-ha’’ and 268.8 cm, respectively) as compared to broad-leaf group (11.1 t-ha’ and 251.1 cm, respectively). A problem with naiTow leafed-cultivars is that the leaves superficially resemble those of marijuana {Cannabis sativa L.). There have been instances where narrow-leafed kenaf plants have been mistaken for marijuana plants. However, there are simple differences between kenaf and marijuana for identification purposes. The marijuana stalks are four-sided without thorns whereas kenaf stalks are generally round and have thorns. A marijuana leaf consists of seven or nine individual leaves joined at a common stem, whereas kenaf leaves are classified as compound leaves with seven lobes (Somers, 1991). We suggest that it may be desirable to grow broad-leafed cultivars at least until kenaf becomes a popular crop and potential confusion can be avoided.

CONCLUSIONS

The main goal of these research efforts was to evaluate the feasibility of kenaf production in Virginia. Results from two years of research indicate that under Virginia conditions, kenaf can be successfully planted at varying row spacings. The fertilizer needs of kenaf seem to be modest, approximately 64 kg- ha ' of nitrogen may be adequate. The P and K content in most soils in Virginia is expected to be adequate for kenaf production. Our results also indicated that up to 1 6 t ha ' dry matter yield can be obtained from kenaf grown as an annual crop. These results indicate that kenaf can be easily produced in Virginia.

ACKNOWLEDGMENT

The seeds for various experiments and details of kenaf cultivars were provided by

C.L. Webber, III. The experiments were conducted by Harbans L. Bhardwaj. Both

authors analyzed the data and prepared the manuscript.

LITERATURE CITED

Bhardwaj, H.L. and C.L. Webber. III. 1994. Seasonal variation in kenaf yield and quality. Proc. 6th International Conference, March 8-10, 1994. New Orleans, LA. International Kenaf Association, P.O. Box 7, Ladonia, TX. p. 150-154.

Bhardwaj, H.L., M. Rangappa, and C.L. Webber, III. 1995. Potential of kenaf as a forage. Proc. of 7th International Kenaf Conference, March 9-10, 1995. Irving, TX. International Kenaf Association, P.O. Box 7, Ladonia, TX. p. 95-102.

Dempsey, J.M. 1975. Fiber Crops. The University Press of Florida, Gainsville.

Nelson, G.H., H.J. Nieschlag, and LA. Wolff 1962. A search for new fiber crops, V. Pulping studies on kenaf TAPPI Journal 45(10): 780-786 (The TAPPI Journal http://www.tappi.org/index.asp?pid=16069&ch=l is published by TAPPI, the technical association for the worldwide pulp, paper, and converting industry).

Rangappa, M., A. A. Hamama, and H.L. Bhardwaj. 2002. Legume and grass cover crops for seedless watermelon production. HortTechnology 12(2):245-249.

SAS. 1996. SAS system for windows, version 6.1 1. SAS Inst., Inc., Cary, NC.

Somers, J.R. 1991 . Is it pot or is it not?. The Maryland Trooper. Maryland Troopers Association, Pikesville, MD.

Wilson, F.D., T.E. Summers, J.F. Joyner, D.W. Fishier, and C.C. Seale. 1965.

120

VIRGINIA JOURNAL OF SCIENCE

'Everglades 41' and 'Everglades 71', two new cultivars of kenaf (Hibiscus cannabinus L.) for the fiber and seed. Florida Agricultural Experiment Station Circular S-168.

Webber, C.L., III and R.E. Bledsoe. 1993. Kenaf: Production, harvesting, and products. In J. Janick and J.E. Simon (eds.), New Crops. John Wiley, New York.

Virginia Journal of Science Volume 56 Number 3 Fall 2005

Fish Consumption Patterns of Populations in Vicinities of Lake Kastoria and Lake Pamvotis, Greece Eugene G. Maurakis,l ’2,3 David V. Grimes, ^ and Dimitra Bobori ^

^ Science Museum of Virginia, 2500 W. Broad St., Riehmond, VA 24542 ^Sehool of Continuing Studies, University of Riehmond, VA 23173 ^School of Environmental Scienee and Policy, George Mason University, Fairfax, VA 22030,

4 Virginia Department of Environmental Quality 4949 A Cox road, Glen Allen, VA 23060 ^Department of Biology, Aristotle University, Thessaloniki, Greece 54006.

ABSTRACT

Objectives are to establish fish consumption patterns of populations in vicinities of two lakes (Kastoria and Pamvotis) in Greece for use in the assessment of risks associated with consumption of fishes in these agri- chemically impaired lakes. Parameters measured were demographics (i.e., gender, age, weight, education level, occupation, residency), freshwater fish eating frequency, species and sizes of fishes consumed, and fish consumption habits [i.e., quantity, parts, and preparation method). All annual mean site- specific consumption rates of the four gender-age class sub-populations surveyed in vicinity of Lake Kastoria (avg. range= 0. 1 03-0.29 kg/day) exceed those of Greece (0.066 kg/day), EU (0.068 kg/day), Spain (0.104 kg/day), Portugal (0.159 kg/day), and the USEPA default value (0.054 kg/day) with two exceptions. Female consumption rates (0.087-0.103 kg/day) of Perea fliiviatilis were below annual consumption rates of Spain and Portugal. Similarly, annual mean site-specific consumption rates of Anguilla anguilla and Cyprinus carpio by male (0. 199-0.210 kg/day) and female adults (0.096- 0.157 kg/day) in vicinity of Eake Pamvotis exceeded those of Greece, EU, Spain, Portugal, and the USEPA default value. Survey results indicate better- educated Greeks to be higher consumers of fish; however, market availability appears to be a stronger determinant of food choice in comparison to health education. All populations in vicinities of both lakes preferred to eat fried fishes with one exception: grilled C. carpio from Lake Pamvotis was preferred by female and male adults.

Keywords: fish consumption, risk assessment, chronic daily intake, Greece

INTRODUCTION

Bioconcentration of environmental contaminants is receiving increased attention from Europe’s general public, for-profit and non-profit institutions, regulatory agencies of countries in the European Community (EC), and the European Environmental Agency (EEA)(Hill, 1999; Millstone et ak, 2000; Tait, 2001; Tait and Quin, 1998; WHO, 1989; UNEP, 1999). Increased concern stems from increased public and

122

VIRGINIA JOURNAL OF SCIENCE

governmental awareness of their impacts on human health, tourism, and other economic issues that effect gross domestic markets (Braxton and Frewer, 1 998; Hites et al., 2004; Miles and Frewer, 1998; World Bank, 1994). Public demand for healthful foods is causing considerable shifts in commodity markets (Millstone et ah, 2000; Tait, 2001; Kafka and Alvensleben, 1997; World Bank, 1992), and demands for clean water has now become a dominant issue in the tourism industry (UNEP, 1999; World Bank, 1994).

In Greece, fish consumption has been increasing since 1 995, concurrent with public awareness that seafood is healthy and can substitute for meat consumption (USDA. 2003). Per capita consumption of fisheries products in Greece is 24 kg (USDA. 2003). With the increased use and exploitation of localized fisheries comes an incremental increase in exposure to local pollution problems (Szucs and Grasselli, 2004; World Bank, 1992). The problem may be particularly acute in highly localized fisheries such as those in lakes, rivers, and natural water dependent aquaculture facilities as they tend to be located in, or impacted by, high yield agriculture areas (Stickney, 1979; Landau, 1992; Stevenson, 2004). Such is the case in Lake Kastoria, Lake Pamvotis, and Lake Vistonis in Greece. For example, we identified 10 pesticides (Atonik, Cobex, Efmathrin=Efmetren/Permethrin, Fusilade, Primextra, Ridomil, Sencor, Targa, Thimet, and Thiodan=Endosulfan)) that were routinely applied prophylatically to agricultural fields draining into Lake Vistonis in 1993, where acute toxicity had been reported to decimate the population of Cyprimis carpio in an aquaculture facility operated by monks (Alpanezos, pers. comm.). Consumption of freshwater fishes from Lake Pamvotis has diminished in recent years because of increased public awareness of the pollution state of the lake (Filos, pers. comm., 2002). Whereas studies in the variability of fish consumption within 10 European countries, nutrient content of foods in Greece, policy analysis on food safety in Europe, and the association of cooking methods to cancer have been conducted (Boulous et al., 1996; Millstone et al., 2000; Rohrmann et al., 2002; Welch et al., 2002), there appear to be no published data on human consumption patterns of freshwater fisheries products for use in calculating risks associated with consumption of freshwater fishes in Greece.

Our objective is to establish fish consumption patterns of populations in vicinities of two lakes (Kastoria and Pamvotis) in Greece for use in assessing risks associated with consumption of fishes in these two agri-chemically impaired lakes.

STUDY AREA

Lake Kastoria (~28 km^; karst basin=304 km^; avg. depth=4 m; max. depth=8.5 m), a eutrophic lake with large concentrations of phytoplankton and mats of submerged aquatic vegetation (Aliakmon River drainage), is located in Kastoria (Macedonia Prefecture) in northwestern Greece (Skoulikidis et al., 1998). The lake, which receives runoff from agricultural operations (e.g., corn, apple orchards, livestock) and a furrier industry, has dropped ~ 1.5 m in the last 5 years, primarily through water withdrawal for agricultural operations (Filos, pers. comm.). Mean physical and nutrient concentrations in the lake are pH (8.2), total phosphorus (39 p-g/1), P-PO4 (31,3 pg/1: max=62.5), N-NO2 (5.8 pg/1; max=19.0), N (15.8 pg/1; max=1011); N-NO3 (22.4 pg/1), and N-NH4 (288 pg/l)(Hadjibiros et al., 1 998). Mean heavy metal concentrations for Pb (3 1 . 1 ppb), Zn (32.8-8 1 .2 ppb), Cu (6.6- 1 9.4 ppb), Cd (0.7 ppb), Hg (0.25 ppb) and As (1 1.1 ppb) have been recorded from Lake Kastoria (Hadjibiros et al., 1998). Main sources of cadmium, copper, lead, and zinc are fertilizer and pesticide residues (Hadjibiros et al., 1998).

FISH CONSUMPTION PATTERNS

123

Lake Pamvotis (22 km^; basin=330 km^; avg. depth=5.5 m; max. depth=l 1 m) is an eutrophic lake beside the city of loaenina (Epiros Prefecture) in western Greece (Skoulikidis et al., 1 998). Mean physical and nutrient concentrations in the lake are pH (8.4), total phosphorus (38 pg/1), P-PO4 (31.7 pg/l: max=91), N-NO2 (2.2 pg/l; max=65.3), N (24.1 pg/1; max=926); N-^N03 (27.3 pg/1), and N--NH4 (62.8 pg/1) (Hadjibiros et ah, 1998). Mean heavy metal concentrations for Zn (33.1 ppb), Cu (5.2 ppb), Cd (1.6 ppb), and As (1,9 ppb) have been reported from Lake Pamvotis (Hadjibiros et al, 1998). Eutrophication is primarily a result of domestic wastewater inputs (Hadjibiros et al., 1998).

MATERIALS AND METHODS

Fish consumption patterns were determined through the use of personal interviews with 90 people selected at random in Kastoria Greece for Lake Kastoria; and 135 people in loannina, Greece for Lake Pamvotis. Fish consumption survey interviews, modeled after guidelines and recommendations in USEPA (1998), were conducted in Greek for Lake Kastoria on June 5, 6, 10, and 17, 2002; and in loannina for Lake Pamvotis, Greece on June 6, 7, 8, 9, 11, 28, 29, 30 and July 1, 2002. Variables recorded included: demographics [i.e., gender, age, weight (kg), education (0=none, 1 =elementary, 2=middle school, 3=E^“2"'^ years of high school, 4=3’"'* and 4^^ years of high school, 5=graduate school), occupation (O^unemployed/retired; l=student; 2=housewife; 3=agri cultural worker/fisherman; 4=blue collar; 5=professional), residency (l=yes, 2=no), domicile distance from lake (m)] and eating statistics [i.e., general marine and freshwater fish eating frequency (days/year); specific freshwater species and sizes (kg) eaten {Perea fluviatilis, Rutihis rutilus for Lake Kastoria, and Anguilla anguiila, Cyprinus carpio, Silurus aristotelis for Lake Pamvotis), specific eating frequency (days/year) per freshwater species from each lake, and fish consumption habits [i.e., quantity (kg), parts (frequency consumption of muscle, skin, bones, head, intestine), and preparation method (fry, grill, boil, head boil, soup frequency). Selection of fish species was based on freshwater fish eating preference, species availability per lake, and ecological feeding type: food chain position (predator, prey); feeding type (i.e., carnivore, herbivore, omnivore), and feeding position (e.g. bottom) as follows: herbivorous bottom feeder=i?. rutilus and C carpio; carnivorous bottom feeder= S. aristotelis; carnivorous pelagic feeder: P. fluviatilis; and carnivorous bottom feeder= A. anguiila.

Four gender-age class groups (male adult > 1 8 yo; female adult > 1 8 yo; male youth <18 yo; female youth <18 yo) conform to those specified in USEPA (1998). Percentages and other proportional measurements were converted into arcsin equivalents to normalize variance prior to statistical tests. Spearman’s correlation analysis (SAS, 2002) was used to identify significant correlations among variables per lake and species. Option NOMISS was employed to eliminate observations with missing values as pairwise correlation matrices may not be nonnegative definite, and the pattern of missing values may bias results (SAS, 2002). Analysis of variance followed by Duncan’s multiple range test (SAS, 2002) was used to test differences in each variable among gender~age class groups by lake and species in order to distinguish group specific consumption patterns from overall consumption patterns.

RESULTS

Lake Kastoria demographics- Mean ages and weights of male (47.3 y; 85.9 kg) and female (48.5 y; 63.5 kg) adults were significantly greater than those of female (12 yo; 37.4 kg) and male (10.4 y; 38.6 kg) youth (Table 1). Mean grade level (1.75) of male

124

VIRGINIA JOURNAL OF SCIENCE

youth was significantly lower than those of male and female adults (range=2.7-2.8; Table 1 ). Occupation level of male adults (3.8) was signifieantly greater than those of female adults (2.8) and male and female children (Table 1). Mean eating days (126.2) of marine and freshwater fishes by male adults were significantly greater than those (range=76.9- 105.8 days) of other interview groups, and negatively correlated with weight (coefficient=-0.3175; p=0.0336; Table 1).

Lake Kastoria Consumption Patterns by Speeies Consumed.

Perea fluviatilis: Annual mean consumption of P. fluviatilis ranged from 0.087 kg/day in female adults to 0.273 kg/day in male adults and (Table 2), and did not vary signifieantly among gender-age groups (F=1.70; p=0.1889). Mean fish size (range=0.32-0.50 kg) and number of days (range= 57.5-73.9) of eating P. fluviatilis did not vary significantly among all gender-age groups (Table 3). However, quantity (x=0.93 kg) of P. fluviatilis consumed per meal by male adults was significantly greater than those ( range x=0.34-0.45 kg) consumed by other gender-age groups (Table 3).

Interviewee weight was significantly positively correlated with age (p<0.0001), education (p=0.0037), occupation (p<0.0001) quantity of P. fluviatilis consumed (p=0.004), and eating skin (p=0.0072) and bones (p=0.02 1 8) of the speeies. Frequency of eating skin was correlated with quantity of P. fluviatilis consumed (p=0.0204). Perea fluviatilis consumption days were correlated positively with frequency of eating skin (p=0.0163) and negatively with eating heads (p=0.0099). Boiling heads of P. fluviatilis was eorrelated with frequency of eating bones (p=0.0136) and soup (p<0. 0001). Fish size and frequency of eating fried P. fluviatilis were negatively eoiTelated (coefficient=-0.4906), whereas fish size and frequency of eating grilled P. fluviatilis were positively correlated (coefficient=0.3580; p=0.0376) indicating that smaller fish were fried and larger ones grilled. Mean percents (65-67) of consumption of fried P. fluviatilis were significantly greater than those of grilled (x range=27-35) and boiled (x range=0-3) in female and male adults (Table 4).

Mean skin eating frequencies (range=82-90 %) in female and male adults were significantly greater than those (x range=0-33 %) in female and male youth (Table 3). Mean frequencies of eating heads (60 %) and bones (40 %) by male adults were significantly greater than those (x range=0-l %) in female and male youth (Table 3). Mean percents (range 33-55) of consumption of fried P. fluviatilis were significantly greater than those of other preparation methods except by female youth (Table 3). All P. fluviatilis were gutted prior to consumption, regardless of preparation method.

Rutilus rutilus: Annual mean consumption of R. rutilus ranged from 0.185 kg/day in female adults to 0.297 kg/day in male children (Table 2), and did not vary significantly among gender-age groups (F=l .68; p=0.1970). Mean quantity (x=0.93 kg) of R. rutilus consumed per meal by male adults was significantly greater than those (x range=030-0.48 kg) consumed by other gender-age groups (Table 5). Mean fish size (range=0.10-0.12 kg) and number of days (range= 77.0-96.7) of eating R. rutilus did not vary signifieantly among all gender-age groups (Table 5).

Interviewee weight (correlation=0.5340; p=0.0028) and occupation (correlation=0.4 178; p=0.0269) were significantly correlated with quantity oiR. rutilus consumed. Age was significantly correlated with eating skin (correlation^O. 36984; p=0.0483) and bones (eorrelation=0.4550; p=0.0131) of R. rutilus. Eating skin was significantly correlated with oceupation (correlation=0.3908; p=0.0398), and eating heads (correlation=0.4761 , p=0.0090) and bones (correlation=0.3864; p=0.0384), negatively eorrelated with grilling R. rutilus (correlation=-0.3864; p=0.0384), and

FISH CONSUMPTION PATTERNS

125

number of days per year eonsuming R. rutilus (correlation=-0.581 1; p=0.0009).

Frequencies of eating skin, heads, and bones (fried or grilled) did not vary significantly among gender-age groups (Table 5). Percent frequencies (x range=67-80 %) of eating fried R. rutilus by all gender-age groups was significantly greater than eating the species either grilled (x range=20-33 %) or boiled (0 %)(Table 5). All R. rutilus were gutted prior to consumption, regardless of preparation method, and none was boiled to make soup.

Lake Pamvotis demographics (based on 135 interviews): Mean ages and weights of male (43.2 y; 81 .9 kg) and female (44.6 y; 68.6 kg) adults were significantly greater than those of female (10.1 y; 40.6 kg) and male (1 1.3 y; 44.4 kg) youth (Table 1). Mean educational levels of female (3.8) and male (3.5) adults were significantly greater than those of male ( 1 .4) and female ( 1 .2) youth (Table 1 ). Occupation levels of female (3.7) and male (3.0) adults were significantly greater than those of female (1.1) and male youth (0.96)(Table 1). Mean distance (1270.8 m) between homes and lakes of male youth were significantly greater than those of adult females (640.2) and males (640.0; F=2.81; p<0.042). Mean eating days (x range=40.5-60.9) of marine and freshwater fishes did not vary significantly among all gender-age groups (F=7.86;

p<0.0001).

Anguilla anguilla: Annual mean consumption of A. anguilla ranged from 0.010 kg/day in female children to 0.199 kg/day in male adults (Table 6), and did not vary significantly among gender-age groups (F=1.31; p=0.3193). Mean number of consumption days (55) for .4. anguilla by male adults was significantly greater than those (x range= 1-6.5) for female and male youth (Table 6).

Age (correlation=0.5360; p=0.0001), weight (correlation=0.3828; p=0.0087), and frequency of eating fried fish (correlation=0.5647; p<0.0001) were correlated with frequency of eating A. anguilla. Frequency of eating soup made from A. anguilla was significantly correlated with boiling heads (correlation=l .0; p<0.0001) eating bones (correlation^l.O; p<0.0001), eating heads (correlation=1.0; p<0.0001).

Quantity, fish size, frequency of eating skin, heads, bones by any preparation method (i.e., fried, grilled, or boiled) of^. anguilla did not vary significantly among gender-age groups (Table 6). Consumption of fried A. anguilla (63 %) was significantly greater than other preparation methods (grilled, 37 %; boiled, 0 %) in female adults (Table 7). Correspondingly, male adults ate fried (48 %) and grilled (48 %) A. anguilla significantly more than boiled (4 %) A. anguilla (Table 7). All A. anguilla were gutted prior to consumption, regardless of preparation method.

Cyprinus carpio: Annual mean consumption of C. carpio ranged from 0.056 kg/day in female children to 0.210 kg/day in male adults (Table 1), and did not vary significantly among gender-age groups (F=l . 17; p=0. 33427). Mean quantity (0.86 kg) of C. carpio consumed per meal by male adults was significantly greater than those (x range=0.2 1-0.25 kg) consumed by female and male youth (Table 8).

Interviewee weight was significantly correlated with age (correlation=0.4768, p=0.0011), education level (correlation=0.4768, p=0.0011), occupation (correlation=0. 0.5868, 0.0011), and quantity of C. carpio consumed

(correlation=0.7959, p<0.0001). Occupation was correlated with eating grilled C. carpio (correlation=0.4041; p=0.0406) and negatively correlated with eating fried C. carpio (correlation=-6645; p=0.0002). Interviewees consuming a greater quantity of C. carpio were more likely to boil C. carpio (correlation=0.6040; p=0.0009), heads of C. carpio (correlation=0.6040; p=0.0009) and eat soup (correlation=0.60 1 8; p=0.0009).

126

VIRGINIA JOURNAL OF SCIENCE

Eating skins of C. carpio was correlated with frequencies of eating heads (correlation=0.91 17; p<0.0001), boiling C. carpio (correlation=0.5449; p=0.0033), boiling heads (correlation=0.5449; p= 0.0033), and eating soup made from boiled flesh and heads of C. carpio (correlation^O.5328; p=0.0042). Eating heads of C. carpio was correlated with boiling flesh of C. carpio (correlation=0.5976; p=0.0010), heads of C. carpio (correlation=0.5976; p=0.0010), and eating soup (correlation=0.5844; p=0.0014). Likewise, eating bones of C. carpio was correlated with boiling flesh (correlation=0.5547; p=0.0027), boiling heads (correlation=0.5547; p=0.0027); and frequency of eating soup made from C carpio (correlation=0.7167; p<0.0001). Grilling C. carpio was inversely correlated with frying the species (correlation— 0.5437; p=0.0034). Grilling was correlated with number of consumption days for C. carpio (correlation=0.4710; p=0.0132).

Percentages of eating grilled C. carpio in female adults (78) and male adults (63) were significantly greater than those for fried (18-28 %) and boiled (4-9 %) in female and male adults (Table 7). Fried C. carpio (70 %) consumed by female youth was significantly greater than that consumed boiled (0 %)(Table 7).

Siiurus aristotelis: Annual mean consumption of S. aristotelis ranged from 0.02 kg/day in male adults to 0.042 kg/day in female adults (Table 1), and did not vary significantly among gender-age groups (F=0.30; p=0.8217).

Interviewee age was significantly correlated with quantity of S. aristotelis consumed (correlation=0.5546; p=0.0258)(Table 9). Eating skin was significantly correlated with eating heads (correlation=0.5092; p=0.0440), grilling (correlation=0. 8783 ;p<0. 0001), and number of consumption days (correlation=0. 7398; p<0.0001) forA aristotelis. Consumers living closer to Lake Pamvotis ate 5. aristotelis on fewer days than those living further away (correlation=-0.3357; p=0.0278).

Mean numbers of consumption days (7.5-1 1.9 days), quantity (0.27-0.50 kg) and fish size (0.61-1.1 kg) of S. aristotelis, parts eaten [skin (38-100 %), head (0-40 %), bones (0-12 %)], and preparation methods [fry (88-100 %), grill (50-100 %)] did not vary significantly among gender-age groups (Table 9). Consumption of fried (x range=69-70 %) S. aristotelis by female and male adults was significantly greater than that for grilled (x range=30-31 %) or boiled (0) S. aristotelis (Table 7). There was no difference in the percent of fried (50) and grilled (50) 5. aristotelis consumed by female or male youth (Table 7). All S. aristotelis were gutted prior to consumption, regardless of preparation method, and none was boiled to make soup.

DISCUSSION

This is the first published study to generate site-specific consumption rates of freshwater fishes in Greece that can be applied as chronic daily intake values (CDI) in determining carcinogenic and non-carcinogenic human health risks associated with the consumption of fish tissues. All annual mean site-specific consumption rates of gender-age populations in vicinity of Lake Kastoria (avg. range= 0.103-0.29 kg/day) exceed those of Greece (0.066 kg/day; USDA, 2003 and Welch et ah, 2002), EU (0.068 kg/day; Leatherhead Food RA, 200 1 ), Spain (0. 1 04 kg/day; Szucs and Grasselli, 2004), Portugal (0.159 kg/day; Szucs and Grasselli, 2004), and the USEPA default value (0.054 kg/day; USEPA, 1991) with two exceptions. Female consumption rates (0.087- 0. 1 03 kg/day) of P. fluviatilis were below annual consumption rates of Spain and Portugal (Table 2). Similarly, annual mean site-specific consumption rates of A. anguilla and C. carpio by male (0.199-0.210 kg/day) and female adults (0.096-0.157 kg/day) in vicinity of Lake Pamvotis exceeded those Greece, EU, Spain, Portugal, and

FISH CONSUMPTION PATTERNS

127

the USEPA default value (USEPA, 1991) (Table 2). Many of our site-speeifie consumption rates for the two lakes rival those of Native Americans in Alaska whose annual mean consumption rate is 0.324 kg/day, six times greater the USEPA default value (TERA, 1999). As there are such high consumption rates for these two different inland populations, and as exposure varies considerably under different circumstances, we concur with the WHO (1999) which strongly encourages responsible authorities in countries to characterize risk on the basis of local measured or predicted site-specific exposure scenarios and not default values such as those available for USEPA models. To date, no default values have been issued by the EEA. Application of site-specific consumption rates as the GDI in risk assessment investigations is warranted as current consumption rates (0.0.66 and 0.068 kg/day) listed for Greece (USDA, 2003 and Welch et al., 2002), and the EU (Leatherhead Food RA, 200 1 ) would underestimate effects of chronic pesticide and other contaminant exposures to local populations utilizing freshwater fishes as a protein source. For example, the EEA (2004) indicates that pesticides are the most common cause of acute and sub-chronic poisonings because of the amounts of pesticides used in comparison with other chemicals, their high toxicity, and inappropriate storage. EEA (2004) did not comment on long-term exposures to pesticides, but indicated that scientific evidence and information concerning actual exposures to chemical substances and their possible health effects is lacking in most European countries, and that the lack of data for health impact assessment poses a big problem. Further, no association between exceedances of EU standards or (WHO) guidelines for pesticides and the incidence of morbidity or mortality has been established, possibly because of the safety margin built into EU standards/WHO guidelines is considerable, and because of the scarcity of appropriate studies (EEA, 2004).

Average consumption of fishes has increased from 0.039 kg/day (Trichopoulou and Lagiou, 1998) to 0.066 kg/day in Greece from 1980-2003 (USDA, 2003 and Welch et al., 2002). The Greek Ministry of Health and Welfare (2003) states that fish and seafood can physiologically substitute meat and eggs, but culinary, practical, and economic constraints dictate a recommendation of about one serving of fish (0.060 kg) per day. Trichopoulou and Lagiou (1998) indicate that fish and seafood availability in Greece decreases with proximity to urban areas, and can be attributed to the immediate availability of sea products in rural costal areas and islands. Our high consumption rates of fishes from both of these inland lakes, where fishes are plentiful and readily available, are consistent with the