Comparison of vehicle mortality following in ovo exposure of Japanese quail (Coturnix japonica) eggs to corn oil, triolein and a fatty acid mix

ABSTRACT

The use of avian egg bioassays for the determination of embryonic mortality and development effects of toxicant exposure is widespread in ecotoxicology. While these studies have a number of experimental limitations to consider, they offer a rapid, cost effective alternative to maternal feeding studies. In preparing to conduct such studies a number of factors must be taken into consideration, including solubility of the toxicant, dissolution solvent, injection site, volume and incubation position. Species-specific requirements for humidity and position should be considered in order to optimize successful incubation with different species. Japanese quail eggs, were injected prior to incubation with 1µl or 5µl of corn oil, triolein or a fatty acid mix, using air cell or albumen injection. Eggs were incubated according to standard poultry practices or in a prone position to determine if there were any differences in hatching success. Hatching success was reduced in eggs that were injected with 5µl and those incubated in a prone position. The highest rate of hatching success was observed for eggs injected with 1µl of the fatty acid mixture through the air cell.

KEYWORDS:

• air cell injections
• albumen injections
• embryonic mortality
• incubation position
• Japanese quail
• vehicle mortality

Introduction

The avian egg bioassay is a useful method for studying the developmental effects of environmental toxicants. This experimental approach has been widely used across chemical classes including steroid hormones, heavy metals, pesticides, industrial chemicals and crude oil to ascertain potential adverse physiological effects in avian embryos (1–20). While feeding trials are more likely to be consistent with environmental exposure, they are expensive, time consuming and can represent a significant toxicological hazard to both research and animal care staff. Although egg injection studies come with their own set of limitations, they are a useful means of studying the effects of embryonic exposure to toxicants. The dose applied to the egg can be more readily controlled.²¹ Toxicants can be applied to the outside of the shell so as to mimic topical exposure either from parental feathers or nest material, such as is commonly seen in crude oil spills and pesticide applications,^22-25 or albumen, air cell and yolk may be dosed, depending on the solubility of the compounds being tested.^{5,6,9,10,14,15,21,26,27} Exposure can also be reliably aimed at developmental stage in order to tease apart early and late effects of exposure²¹. Potential confounders such as laying order, which can result in intra- and/or inter-clutch variation of some contaminants,^28-30 are removed. Additionally, factors such as parental behavior, parental age and quality, weather and even hypoxia, that can affect toxic endpoints in field studies, can be controlled for in the laboratory.

Toxicant delivery is an important aspect of egg injection studies. Maternal deposition of lipophilic compounds such as polycyclic aromatic hydrocarbons (PAHs) occurs primarily in egg yolk, while compounds such as methyl mercury are distributed in albumen and yolk.^9,31 In order to closely mimic ‘natural’ embryonic exposure the simplest solution would be to inject the compound of interest where it is ‘expected’ to occur in the egg based on solubility and available measurements. However, this raises the question of what effect puncturing membranes has on diffusion within the egg. In addition, the solubility of the compounds in water or lipid phase direct primary localization into the egg yolk or albumen

Natural transfer of maternally deposited PCBs (polychlorobiphenyls) to the embryo via yolk absorption is slow during the first 70–85% of development,³² with wild bird hatchlings having more than half of their maternally deposited PCB present in yolk.^33-35 Puncture of egg and yolk membranes to deliver toxicant to its site of maternal deposition in egg injection studies may alter these processes sufficiently such that it changes developmental exposure patterns or rates. Such possible confounding variables mean that often a balance must be struck between effective delivery of the toxicant to the egg with the least amount of disturbance to the embryo and other egg components.

Egg injection studies have used air cell, albumen and yolk injections. The least invasive means of injection of the toxicant is into the air cell; however it is not without limitations. Polybrominated diphenyl ethers (PBDEs) injected into the air cell showed slow uptake into the egg,²⁷ while methyl mercury delivered by air cell injection readily diffused into the eggs of 4 avian species.⁹ These differences in progression of toxicants into the egg point to several factors, including solubility of the toxicant in its vehicle carrier, ability of that vehicle to diffuse across membranes, and chemical stability.

Foremost in the determination of the vehicle to be used is the solubility of the toxicant. For lipophilic compounds such as PCBs vehicles can be divided into 3 categories: organic solvents such as DMSO (dimethyl sulfoxide) and triolein, plant oils such as corn and safflower oils, and emulsions.^12,19,36-46 While only microliter volumes of organic solvents are used as egg injection vehicles, the possibility remains that while they may not alter overall toxicity, they may disrupt other physiological processes within the developing embryo. The same can be suggested of the plant oils. Not only do plant oils contain different ratios of fatty acids to eggs they also contain plant sterols that can reduce cholesterol uptake,^47-49 which could be detrimental to a developing embryo.

Injection volume can also alter the effectiveness of egg injections. Too little may result in re-crystallization of the toxicant leading to reduced delivery, while too much may go so far as to cause hypoxia.^42,43 Vehicle volume may also be limited by the injection site used. Air cells may have less capacity to absorb without leakage or wicking along the shell compared to yolk or albumen, so volume must be adjusted to prevent leakage or wicking along the shell.

For standard poultry practices eggs are incubated ‘upright’ with the air cell or blunt end of the egg pointing upwards. Racks are tilted 60° every hour. While this practice maximizes space for production purposes it may not mimic the natural environment, and thus also acts to alter embryonic exposure patterns. When PCB 126 (3,3′,4,4′,5-pentachlorobiphenyl) was introduced into the air cell of mallard eggs there was up to a 9-fold decrease in toxicity if the eggs were incubated in a ‘horizontal’ position and rotated by 180° every hour compared to eggs incubated according to standard practices.¹² Egg rotation and incubation position are likely to be quite variable among bird species, depending on the type of nest, number of eggs and parental attendance at the nest. In further developing the ‘injection model’ for a given toxicant this factor may also need to be considered. Early diffusion of a toxicant across the air cell directly into the embryo may be effectively increased by the closer proximity of the embryo to the source of the toxicant. In the vertical position the early stage embryo is located directly beneath the air cell, so the potential for early passive diffusion is quite high. In the case of horizontally incubated eggs there is an increased diffusion distance between the air cell and the embryo,¹² which may be sufficient to reduce such early exposure.

The purpose of this study was to investigate differences in hatching success of fertile Japanese quail eggs treated prior to incubation with 3 different lipophilic vehicles, 2 injection volumes, air cell or albumen injections, and vertical or horizontal incubation positions. Although this quail colony is maintained on a stimulatory photoperiod (15L:9D) and constant environment year round, there is some indication that birds still respond to subtle seasonal cues. We also compared experiments conducted over the year to ascertain if there are seasonal differences in optimal hatching success. Finally, aging also impacts the hatching success of eggs from quail colonies, with declining reproductive success. We also tested the best seasonal timing to achieve optimal hatching success in this species. Reproductive success in the Japanese quail begins to decline as early as 56 weeks.⁵⁰ Accordingly all pairs that produced fertile eggs used in this study were under 12 months of age.

Results

Hatching success of untreated Japanese quail eggs (Table 1), incubated in either vertical (78.26%) or horizontal (79.17%) positions were similar and within expected parameters for the University of Maryland random bred colony. Hatching success decreased with all treatments, but the stepwise log-linear model detected only 7 significant decreases (Table 2; all p < 0.03). Eggs injected with 5µl of any vehicle and eggs incubated horizontally, regardless of vehicle or injection volume, were found to have significantly lower (p = 0.0007 and 0.0001, respectively) hatching success (Table 2). There was an effect of vehicle on hatching success, with significantly lower (p = 0182) hatching success detected for eggs treated with corn oil or triolein as compared to no injection, sham injected or the fatty acid mixture vehicle treated eggs (Table 2). Further, there was a statistically significant (p = 0.0135) difference between the untreated and sham injected eggs and those treated with the fatty acid mixture vehicle (Table 2); however, this is likely confounded by eggs within that group treated with 5µl of vehicle and those incubated horizontally. Eggs injected with 1µl of the fatty acid mixture through the air cell, then incubated in an upright position had hatching success of 80% (Table 1), which is almost identical to hatching success of the control groups (untreated and shams). Further, hatching success of eggs treated with 1µl of corn oil through air cell injections, and incubated vertically was 75% (Table 1), similar to controls. Albumen injections were also found to result in significantly lower hatching success of eggs than air cell injected eggs (Table 2). There was an interaction of volume and vehicle with eggs treated with triolein and corn oil with significantly lower (p = 0.0279) hatching success than untreated, sham injected and fatty acid mixture treated eggs (Table 2). Further there was also an interaction of incubation position and injection site with albumen injected eggs having significantly lower (p = 0.0215) hatching success than air cell injected or uninjected eggs.

Table 1. Percentage (%) hatching success of Japanese quail eggs injected through the air cell or albumen with 0.1 or 0.5µl per gram of egg of corn oil, triolein or the fatty acid mix, then incubated in the ‘vertical’ or ‘horizontal’ position. All eggs are rotated through 60° every hour.

			injection volume (µl/g egg)
			0.5		0.1
vehicle	injection site	incubation position	% hatch	n	% hatch	n
fatty acid mix	air cell	vertical	60.87	23	80.00	25
		horizontal	54.55	22	63.64	22
	albumen	vertical	69.57	23	73.91	23
		horizontal	34.78	23	60.00	25
corn oil	air cell	vertical	65.22	23	75.00	24
		horizontal	40.91	22	66.67	24
	albumen	vertical	52.17	23	62.50	24
		horizontal	13.04	23	47.83	23
triolein	air cell	vertical	34.78	23	60.87	23
		horizontal	40.00	25	64.00	25
	albumen	vertical	39.13	23	60.00	25
		horizontal	13.04	23	39.13	23
shams	air cell	vertical	76.60	47
		horizontal	78.72	47
	albumen	vertical	70.21	47
		horizontal	63.27	49
untreated	N/A	vertical	78.26	46
		horizontal	79.17	48

Table 2. Step-wise log-linear regression results of hatching success following egg injections at ED0. Results are shown for terms and interactions for which p < 0.05.

treatment and interactions	df	ChiSquare	p value
injection volume	1	17.058	<0.0001
incubation position	1	11.488	0.0007
vehicle [triolein & corn oil – fatty acid & sham & untreated]	1	5.579	0.0182
vehicle [fatty acid – sham & untreated]	1	6.097	0.0135
injection site [albumen – air cell & none]	1	12.802	0.0003
volume * vehicle [triolein & corn oil – fatty acid & sham & untreated]	1	4.833	0.0279
incubation position * injection site [albumen – air air [air cell & none]	1	5.286	0.0215

As described in the Methods, the University of Maryland random bred colony is maintained on controlled environmental conditions with a stimulatory photoperiod to maintain production of fertile eggs throughout the year. However, anecdotally there appears to be a reduction in egg quality/hatching success in the latter half of the year (from July until January) during the time that the birds would be photoregressed in the wild. To test if time of year affects egg quality and consequently impacting egg injection studies, the groups with the highest hatching success (fatty acid and corn oil vehicle incubated upright following 1µl air cell injections) were repeated in mid-July, and compared to experiments shown in Table 1 that were completed in March and April. Hatching success (Table 3) was greater in eggs injected earlier in the year, compared to eggs injected after the longest day of the year. Nominal logistic fit showed that there was a significant difference in hatching success with time of year (df = 1; Chi square = 9.15; p = 0.0025); however there was no difference between treatment groups (df = 3; Chi square = 0.132; p = 0.988) or with treatment and time of year (df = 3; Chi square = 0.232; p = 0.972).

Table 3. Effect of time of year on hatching success of vehicle, sham and untreated Japanese quail eggs from the University of Maryland, College Park, random bred colony. All laying pairs were less than 12 months of age, eggs were incubated according to standard poultry practices and were injected with 0.1µl/g of egg of vehicle into the air cell.

	Jun 21 –Dec 21		Dec 21-Jun 21
treatment	% hatch	n	% hatch	n
corn oil	56.25	16	75.00	24
fatty acid mix	56.25	16	80.00	25
untreated	56.25	16	78.26	46
sham	58.82	17	76.60	47

Discussion

The avian egg bioassay is only as good as its vehicle controls. Stated another way, the ability to discern effects of test compounds in an avian egg bioassay depends on consistency of response, both in the vehicle control and untreated groups to provide a sound basis for comparison to effects of a test compound. Inadvertent embryonic mortality in response to the vehicle will reduce the detection sensitivity of significant effects, increase the sample size required for sufficient statistical power, and impact accurate determinations of the dose-response relationship. Previous egg injection studies have shown that hatching success can vary between species for to a number of variables including, day of injection, injection vehicle, volume injected and site of injection.^{9,10,42,43,51} Optimal incubation procedures and conditions for embryonic viability with consideration of species-specific developmental requirements^9,12 have received less attention. The combination of incubation practices used in domestic poultry and husbandry information pertinent to a variety of species has provided a basis for approaches used across a variety of species. However, few studies have directly compared these key variables that occur in egg injection studies, namely vehicle, injection volume, site of injection with placement of the treatment compound within the egg matrix, and position for incubation. Many of these variables have been determined by the compounds used and the avian species studied. Nonetheless, the issue of minimizing the impact from vehicle(s) has not received as much attention. The current study sought to determine the best conditions for injection and incubation of Japanese quail eggs in order to optimize viability.

As predicted, incubation in an upright position, with the air cell pointing upwards was the optimal incubation position for this species. Japanese quail incubate their eggs in a cup shaped nest of straw/grass-like materials, in which most of the eggs are aligned in more or less a vertical position.⁵² Nest structure and clutch size are variable between bird species, so differences would be expected in incubation position and turning rates between species, which in turn could influence the selection of an egg injection protocol. For mallard eggs there were are no statistical differences in vehicle injected and untreated mallard eggs regardless of incubation position.^9,12 However, this was not the case with Japanese quail eggs for which hatching success was reduced in all vehicle injected eggs incubated in a ‘horizontal’ position with hourly rotation of 60°, but not in sham injected or untreated eggs. Our incubation paradigm was not identical to that used by Heinz et al., (2006)⁹ or McKernan et al. (2007),¹² so it is possible that the lack of complete 180° turning is responsible for the observed differences in vehicle treated eggs. It is also possible that the natural incubation ‘procedure’ may also play a role in inter-species differences in hatching success relative to incubation position.

Based on our data, injection volume is a critical factor in the successful incubation of air cell injected eggs. Not only is space limited such that solution loss can occur by wicking on the surface of the egg, but it is possible that too high a volume can affect early embryonic development if the vehicle prevents adequate respiration.^42,43,51 Further, large vehicle injection volumes could potentially impact nutrient absorption by the embryo at hatch and early post-hatch, during the time of maximal use of the remaining yolk. Finally, relatively large injection volumes could damage shell integrity, allowing potential bacterial or fungal infection. Vehicle volumes tested in this study were low compared to other studies, which have used 1–5µl per gram of egg^9,10,42; however, there were still reductions in hatching success in the groups that received 0.5µl per gram of egg compared to those that received 0.1µl per gram of egg. Volumes up to 2.0µl per gram of egg for mallards or double-crested cormorants, did not reduce hatching success in vehicle treated controls.^9,12 This may be associated with injection on embryonic day (ED) 4, which Heinz and colleagues (2006)⁹ suggested as an optimal stage for injection. They suggested that egg injection at earlier stages may be embryotoxic or result in hypoxia/anoxia.^42,43,51,53

As described in the Methods section the University of Maryland random bred quail colony exhibits some signs of maintaining natural photoperiodicity, despite maintenance of long day length. The increase in control mortality in our study showed that there does appear to be evidence for reduced reproductive success during the second half of the calendar year. Vehicle injected eggs during what would be the naturally photoregressed time of year showed lower hatching success compared to earlier in the calendar year. While this does not specifically affect these experiments, other than to ensure that egg injection studies for this colony only occur during the normally photostimulated time of the year, it should be used as a potential caution for other species when egg collections may run into sub-optimal periods with respect to egg production. These scenarios could include periods of poor nutrition, which can influence eggshell thickness and nutrient content of the egg, and less than optimal temperatures and weather conditions for incubation.

From an experimental perspective it is also important to determine if air cell and/or albumen injections can be used to mimic embryonic uptake of maternally transferred toxicants. In particular as injected concentrations of toxicants are increased, membrane saturation could become a rate-limiting factor for diffusion into the egg via the air cell. Further, embryonic exposure to lipophilic compounds injected into the more hydrophilic albumen could result in alterations in developmental exposure patterns relative to maternally deposited compounds. Our study showed the albumen injections increased embryonic mortality due to vehicle injection; however, there is insufficient evidence to determine if this was due to exposure to the vehicle or disturbance to homeostatic processes within the egg through membrane ruptures caused by the injection process.

Embryonic exposure to lipophilic compounds such as PCBs via maternal deposition is primarily by absorption of the yolk.^21,54 Early in development, the albumen makes up approximately 2 thirds of the egg volume and meets most of the energetic requirements of the developing embryo.⁵⁵ This results in slow natural transfer of nutrients, hormones and exogenous lipophilic compounds to the embryo via yolk absorption during the first 70–85% of development, as observed by de Roode and van den Brink (2002)³² following PCB yolk injections and by Bargar et al. (2001)⁵⁶ in a chicken feeding trial. The proportion of PCB reaching the embryo changed from 2% at ED 13 to 18% at ED 19³² and 0.8% at ED 13 to 15% at ED 18.⁵⁷

Under normal developmental conditions the embryo would have less exposure during early embryonic development to exogenous compounds localized in the yolk than they would compared to albumen soluble compounds. Thus, albumen injections for egg bioassays could expose embryos during earlier phases of development to higher concentrations of toxicants than would naturally occur through maternal deposition. While active uptake of the PCB from yolk or albumen during development would contribute to a greater extent of the overall absorption, passive diffusion into the embryo during initial stages of development should not be excluded as a mechanism of exposure. The lipophilic properties of PCBs make them excellent for diffusing across membranes, including skin,⁵⁸ so it is possible that maternally deposited PCBs behave in a similar fashion. Given that quail eggs, like other commercial poultry, are incubated with the air cell at the top, such that the embryo is somewhat ‘sandwiched’ between the air cell and yolk, the ability of the PCBs to reach the embryo is primarily dependent on the permeability of the egg and yolk membranes. In this sense air cell injections may mimic the early developmental exposure of embryos to exogenous lipophilic compounds.

Currently, there are only a few studies that provide insight into the partitioning of PCBs within egg compartments that can provide compelling evidence for air cell injection over albumen injection, or vice versa. A study on PDBEs in avian eggs²⁷ showed that air cell injections of these compounds resulted in diffusion of up to 29% of penta-BDEs into the egg at pipping. Also, a further study on diffusion of radiolabeled [¹⁴C]-PCB 77 from air cell injections prior to incubation in Japanese quail (Dean et al., in prep) showed that after 24 h up to 2% of the dose had reached the yolk.

This is in agreement with yolk and albumen usage by developing embryos. Early in development, the embryo primarily absorbs and uses albumen. In Japanese quail, yolk cells increase in size until ED 5. By ED 10, yolk cell membranes have become quite fragile and easily broken, and by ED 12, the remaining albumen has merged to the yolk sac.⁵⁵ It is during the second half of incubation in precocial species that the yolk also serves as a depot for the steroid hormones produced by the embryo (16, 59). Moreover, potential effects of pesticides and other contaminants may act synergistically or in opposition to the role of endogenous steroids in development of critical physiological systems.^14,16,21,46

Taken together with the results of the current study, administration of the smallest quantity possible of vehicle needed to dissolve the test substance, treating by injection into the air cell, and incubation upright for quail eggs optimize embryonic viability and reduce vehicle-related variability. Further, embryonic exposure increases throughout incubation with the highest potential exposure during the last third of incubation and continued exposure post hatch. This means that lipophilic compound(s) administered into the air cell will have the opportunity to migrate into the yolk and be there in relatively high concentrations during the time of maximal embryonic exposure. While there are differences across avian species in the degree of pre-hatch development, percentage of the egg that is yolk versus albumen, and composition of both egg components, the developmental processes are similar. In the first stages of development organs and systems are being created, which utilizes the high protein content of the albumen; as development progresses there is a switch to using fats as an energy source. So choosing the type of vehicle carrier, volume and injection site should be considered carefully for every species and toxicant of interest. Egg injection studies cannot precisely mimic maternal transfer, but they should be designed according to the species and endpoints of interest.

Materials and methods

Animals

Japanese quail (Coturnix coturnix japonica) from the University of Maryland random bred colony were used for this study. Animal housing and care were under Institutional Animal Care and Use Committee approved protocols. The photoperiod was stimulatory to maintain reproduction (15L:8D), with temperature and humidity controlled throughout the year. Pairs were maintained in appropriately sized cages with an egg roll; feed (https://www.purinamills.com/game-feed/products/game-bird/purina-game-bird-layena-etts/) and water were available ad libitum. Each banded bird was placed in a numbered cage and monitored according to animal care protocols.

Egg injection

Eggs were collected for 2–3 consecutive days from male-female pairs and labeled within 1 cm of the air cell using a soft (8B) pencil (http://www.fabercastell.com/products/pencils/GraphitepencilCASTELL90008B/119008). Following collection, eggs were assessed for damage, then stored under plastic wrap at approximately 10°C until the day of injection. Prior to injection eggs were weighed and randomly assigned to groups. Average egg weight was approximately 9.65 + 0.6 g (mean + SD) across all parts of the study

Eggs were injected prior to incubation on embryonic day 0 (ED0). Each egg was wiped with an alcohol swab, and then a 2–3 mm hole was drilled through the shell into the air cell or albumen using an alcohol sterilized Dremel (https://www.dremel.com/en-us/Tools/Pages/ToolDetail.aspx?pid=8100#.V3VWko7lJF8) with a diamond wheel tapered bit (https://www.dremel.com/en-us/Accessories/Pages/ProductDetail.aspx?pid= 7150). Air cells were visualized with a fiber optic lamp, and holes were drilled in the center. Albumen injections were made with approximately 1.5 cm from the bottom or air cell end of the egg. Eggs were dosed with either 1µl or 5µl of solution into air cell or albumen. Air cell injected eggs were checked to ensure that the air cell was not punctured. Each egg was checked for wicking of the solution. If this occurred, the egg was replaced with another of similar size, and the substitution recorded. Eggs were sealed with low melting point paraffin wax. Sham injected eggs were drilled and then sealed with paraffin wax.

To test if time of year affects egg quality and consequently could impact egg injection studies, the fatty acid and corn oil vehicle groups incubated upright following 1µl air cell injections were repeated in mid-July, compared to experiments, which were completed in March and April.

Incubation

Eggs were incubated under standard poultry conditions in a Georgia Quail Farm (GQF) digital incubator (http://www.gqfmfg.com/1500-series/1500-digital-professional/) incubator set at 38°C and 55% humidity for 14 d. As a variation on the protocol used by McKernan et al. (2007) a subset of eggs incubated ‘horizontally’. The standard GQF incubator cannot rotate eggs by 180°, instead eggs were simply laid on their sides in the tray and exposed to the standard 60° rotation per hour as all other eggs. While this does not mimic previous experiments it is likely to represent a plausible nest scenario.

Eggs were candled every 3 d to determine viability. Dead eggs were removed from the incubator. Following each candling racks were rotated between shelves to ensure even exposure to random variations in temperature, humidity and light.

On day 14 of incubation the eggs rotation was ceased and humidity was increased to 65% for hatching. On day 16 (following candling) eggs were transferred to hatching compartments.

Dosing solutions

Three different vehicles were tested: corn oil (http://www.sigmaaldrich.com/catalog/product/supelco/47112u?lang=en&region=CA); triolein (http://www.sigmaaldrich.com/catalog/product/sial/y0001113?lang=en&region=CA) and a fatty acid mix prepared using the 4 most common fatty acids found in Japanese quail eggs (60), while still maintaining a mixture that was liquid at incubation temperature (38°C). The fatty acid mix was composed of palmitic, stearic, oleic and linoleic acids (http://www.nu-chekprep.com/su.htm) in a ratio of 10:5:55:30. All solutions were filter sterilized with a 0.2µm syringe-driven filter unit before use.

Statistical analyses

Stepwise log-linear model was fitted to hatching success against all 4 test parameters (injection volume, injection site, vehicle and incubation position) and their interaction effects (61, 62). The threshold for including an effect in the model was a p = 0.05 and the best fit model was determined using Akaike Information Criteria (AIC). conducted using JMP 8.0.1 (SAS, Cary, NC). Contingency tests were used to determine if there were differences in hatching success between groups based on time of year when injections were performed.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgments

The authors wish to thank Kara Duffy and Nichola Thompson for their valuable technical assistance, as well as Gary Heinz and the reviewers for their helpful insights.

Funding

This work was supported by the U.S. Fish and Wildlife and the Hudson River Natural Resource Trustees. The conclusions and opinions presented here are those of the authors and do not represent the official position of any of the funding agencies, the Hudson River Trustees, or the United States of America.