Our Analysis of the Electrolyte study
Twenty-one (21) Ruby-throated hummingbirds (wild captured) were measured for mass and cloacal fluid content during an 11 hour analysis (2 analysis per animal).
We feel the study was poorly conceived and implemented.
1) Inappropriate "Outlier" Exclusion
In this study, seven (7) birds were excluded on the basis of a 10+% change in mass from the initial weighing of the test subjects. The authors of the study rationalized that a 10% change in body mass was not possible to be caused by the electrolyte differences in the five (5) different nectar diets. The authors offered no basis for this assumption and excluded these data points without rational cause or analysis.
They did not utilize Peirce's criterion, nor Dixon's Q-test, nor Chauvenet's criterion, nor the ROUT test, nor Grubb's test to determine if this was a valid course of action. Dixon's Q-test would perhaps be the most appropriate test for this sample sizes (between 3 to 25), presuming the data are normally distributed.
Despite the absence of the use of an appropriate test to determine the validity of excluding any individual data point, the exclusion of such a large percentage of the dataset (33%) is also unacceptable. The data exclusion is particularly inappropriate, since the readout that was used for the exclusion (mass) was one of the critical readouts used to infer that their nectar would lead to an "improved hummingbird condition over sugar-water". However, the authors acknowledge that this "improved condition" would only be in the setting of a restriction from access to insects, which is not likely to occur in wild hummingbird populations.
The authors also clearly admit that minimal consumption of insects and the consumption of wild nectar is more than capable of satisfying the daily requirements of electrolytes in the hummingbird's diet. Making the claim of benefit all the more suspect.
They rationalize the removal of these data points by stating that the big swings in mass of the birds was due to differences in "gastrointestinal fill" in recently caught wild birds, without any data to support that contention. To avoid this complication in future studies, the authors of the study suggest that future studies should "allow substantial absorption or passage of the diet most recently consumed" by modifying the protocol such that "the birds should be held for 15 minutes" prior to continuing the study. However, the Methods Section of this study clearly states that the birds were studied after, "an initial period of acclimation to captivity"; indicating that the protocol already included this stipulation.
Unintended researcher bias can’t be excluded, and would be virtually guaranteed by allowing for a removal of 33% of any dataset. Without inappropriately excluding 33% of the data, there was no effect on the change in the mass of the birds when offering them any of the diet preparations, even when offered a diet that was free of electrolytes (sucrose solution only).
Even when inappropriately excluding 33% of their data, the extent of variance in their data (indicated by the error bars on the bar graph in Figure 1) revealed that there still was no significant difference between any of the electrolyte solutions compared to the sugar water solution. As a result, utilizing the data indicating the change in mass of the birds would not allow any rational researcher to determine any benefit nor detriment of supplementing commercial hummingbird nectar with electrolytes.
2) The critical parameters in the diet that were being tested were not independently varied.
Any analysis of one electrolyte should carefully control for the concentration of the other critical electrolytes by holding them constant.
The potassium levels were changed in each of the 5 nectar diet's that were offered to the birds (0, 2.81, 4.88, 14.4, and 31.6 mM K+), while changing the sodium and chloride concentrations at the same time. The sodium and chloride levels for each of the 5 nectar solutions were not held constant while the potassium levels were varied. Therefore, any differences in experimental outcome would not be capable of being ascribed to any individual electrolyte.
In addition, there was only a single nectar solution that contained chloride ions with no comparator group; and the concentrations of sodium jumped 10 fold in the middle of the titration range (from .942 mM or 9.77 mM) with no rational basis for selecting those particular concentrations.
Inferring any correlational or causational effect of electrolyte levels on the health of the birds from the experimental plan within this study is not possible.
As the authors point out, the cloacal release of electrolytes "tracks" with the dietary electrolyte levels; which is the only conclusion that can be gleaned from this experiment and which is consistent with what has been shown in the literature (Calder).
The data provided in this study is not sufficient to allow the authors to conclude that any particular electrolyte solution would be of benefit to the birds. In spite of the limitations of the study, the authors suggest that 14. mM potassium, 10. mM sodium, and 3. mM chloride should be used as a supplement in a commercial nectar.
However, their electrolyte levels suggested actually fall within the levels observed in natural hummingbird nectar (e.g. mean K+ Delphinium nectar= 50. mM, mean K+ Ipmopsis nectar=1.75 mM; Calder and Hiebert, 1983).
From American Ornithologist Union Annual Meeting
Presented in Jacksonville, FL July 2011
Electrolyte and Water Balance in the Ruby-Throated Hummingbird
The liquid diets of hummingbirds can present a physiological challenge because their kidneys display obligatory losses of electrolytes. Thus consumption of human-provided sugar-water that is free of electrolytes may leave birds in negative water balance. The goal of this study was to determine the minimal concentration of sodium and potassium required by Ruby-throated hummingbirds. Individual birds (N=21) were isolated, fed one of 4 diets with added potassium, sodium, and chloride, and were monitored for water and electrolyte balance. Birds maintained body mass on the electrolyte diets but lost mass when consuming sugar-water. Bird consuming < 5 mmol/L of potassium and < 1mmol/L of sodium displayed a net loss of these elements suggesting that these animal were in electrolyte deficit. Birds consuming 32 mmol/L of potassium and 11 mmol/L of sodium excreted excess electrolytes suggesting that requirements had been exceeded. These results indicate that sugar-water likely contributes to electrolyte and water imbalance in hummingbirds and that this problem can be resolved with relatively small supplements of sodium and potassium to artificial nectars.
From Birding Business December 2011
This study examined electrolyte and water balance in the Ruby-throated Hummingbird (Archilochus colubris), the most common visitor to backyard hummingbird feeders in the Mid-West and East.
Nectarivorous hummingbirds consume up to 5 times their body mass in nectar each day, but ironically this exposes them to probable dehydration when consuming solutions that do not contain electrolytes. Hummingbirds quickly process excess water while drinking, which serves to return body mass to pre-feeding status and reduces the ill-effects of excess circulating water. The hummingbird kidney experiences obligatory loss of electrolytes, so when hummers do not consume electrolytes in their diet they will experience a net loss of electrolytes. When the amount of electrolytes in the body drops the concentration of water will also drop as the body acts to maintain a constant electrolyte concentration. This can lead to dehydration and weight loss.
Mario and his team are currently working on more detailed analyses of bird food and nutrition which will be available for publication at a later date. We'll update you on those results as they come in. In the meantime research continues at various other bird food packers, and we'll pass along the detail as it comes to us.