Conflict/competing interest: Aki Kawasaki has received financial compensation from Bayer SpA, Milano Italy as a consultant. None of the authors have a proprietary or commercial interest in any of instrumentation, technology or products mentioned in this paper.
Funding sources: Aki Kawasaki is a recipient of a research grant from Open Eyes Foundation and Loterie Romande, Lausanne, Switzerland.
Aki Kawasaki, MD, Hôpital Ophtalmique Jules Gonin, Avenue de France 15, 1004 Lausanne, Switzerland. Email: email@example.com
Background: In patients with outer retinal degeneration, a differential pupil response to long wavelength (red) versus short wavelength (blue) light stimulation has been previously observed. The goal of this study was to quantify differences in the pupillary re-dilation following exposure to red versus blue light in patients with outer retinal disease and compare them with patients with optic neuropathy and with healthy subjects.
Design: Prospective comparative cohort study.
Participants: Twenty-three patients with outer retinal disease, 13 patients with optic neuropathy and 14 normal subjects.
Methods: Subjects were tested using continuous red and blue light stimulation at three intensities (1, 10 and 100 cd/m2) for 13 s per intensity. Pupillary re-dilation dynamics following the brightest intensity was analysed and compared between the three groups.
Main Outcome Measures: The parameters of pupil re-dilation used in this study were: time to recover 90% of baseline size; mean pupil size at early and late phases of re-dilation; and differential re-dilation time for blue versus red light.
Results: Patients with outer retinal disease showed a pupil that tended to stay smaller after light termination and thus had a longer time to recovery. The differential re-dilation time was significantly greater in patients with outer retinal disease (median = 28.0 s, P < 0.0001) compared with controls and patients with optic neuropathy.
Conclusions: A differential response of pupil re-dilation following red versus blue light stimulation is present in patients with outer retinal disease but is not found in normal eyes or among patients with visual loss from optic neuropathy.
The melanopsin-expressing retinal ganglion cells (MGCs) are a newly described subset of retinal ganglion cells and serve as the principal neural conduit of retinal afferent pupillomotor input for the pupil light reflex.1–4 The activity of MGCs is modulated by both extrinsic and intrinsic light signalling on these cells.5–7 Extrinsic signalling derives from rods and cones and is relayed trans-synaptically to MGCs via bipolar cells. Intrinsic signalling from melanopsin-mediated phototransduction is a unique feature of these ganglion cells and can initiate direct depolarization of the MGC with generation of fast action potentials in the absence of extrinsic rod and cone input.1,6,7 Under physiological conditions, the afferent pupillomotor input to the pupil light reflex represents the summed influence of rod, cone and melanopsin activation to a light stimulus.5–8
Single-cell recordings have demonstrated that narrow bandwidth light at selected wavelengths and intensities can be used to preferentially activate one photoreceptive input (rod or cone or melanopsin) to the MGC6. Kardon et al. defined a clinical protocol using coloured light stimuli intended to bias afferent pupillomotor input in favour of rods, or cones or melanopsin activation and investigated the pupil responses of patients with retinitis pigmentosa.9,10 The authors found that, as expected, pupillary contraction to stimulus conditions weighted to rod activity and to cone activity was reduced in patients with retinitis pigmentosa as compared with those obtained from normal eyes.10 And yet, in patients with severe visual loss and extinguished electroretinograms (ERGs), pupil responses derived primarily from outer photoreceptor input were still easily evoked and recordable.10 An unexpected observation from the pupillographic tracings, however, was that the patients with retinopathy, but not subjects with normal eyes, demonstrated a slow and delayed pupillary re-dilation following bright blue light exposure but not after bright red light exposure. The authors proffered that ‘sustained pupil contraction following blue light offset compared with red light offset may provide additional diagnostic information correlating to the status of outer photoreceptor disease’.
This study was undertaken specifically to address the question of differential pupillary re-dilation dynamics for blue light versus red light among patients with outer retinal disease. The primary goal of this study was to identify objective parameters that describe and quantify pupillary re-dilation and to compare the results from patients with outer retinal disease to patients with optic neuropathy and to subjects with healthy eyes. In addition, this study examined whether the presence of a slow and delayed re-dilation was related to the degree of outer photoreceptor dysfunction, as assessed by ERG and perimetry.
The study was conducted according to the tenets of the Declaration of Helsinki and received authorization from the local ethical committee. All subjects provided oral and written informed consent for study participation.
Patients with a diagnosis of outer retinal disease or optic neuropathy were recruited from the neuro-ophthalmology clinic at the Hôpital Ophtalmique Jules Gonin. Inclusion criteria for patients with outer retinal disease were, based on chart review, an abnormal fundus appearance consistent with outer retinal degenerative disease and an abnormal scotopic and/or photopic full-field ERG, which was performed according to the International Society for Clinical Electrophysiology of Vision standards and recommendations.11,12 None of the patients with outer retinal disease had history of or evidence of another retinal or optic nerve disorder other than the primary diagnosis. An OCT (OCT Stratus 3000, Carl Zeiss Meditec Inc., Dublin, CA, USA) was performed in all patients to ensure that the average nerve fibre layer thickness was in the normal range, that is, there was no evidence of severe optic nerve atrophy. Inclusion criteria for patients with optic neuropathy were unilateral or bilateral optic nerve dysfunction with visual field defect and presence of a relative afferent pupillary defect in the absence of other ocular pathology. No patient had a systemic disease known to cause retinopathy. ERG was not performed in these patients as it is not part of the routine evaluation of patients with optic neuropathy. Patients with arteritic optic neuropathy were excluded because of potential concomitant retinal dysfunction.
Healthy control subjects were recruited from hospital flyers. All participants had an ophthalmological examination with objective and subjective refraction, colour vision testing (Ishihara plates, Kanehara & Co., Ltd., Tokyo, Japan), slit lamp biomicroscopy, direct fundus examination, fundus photography (VISUCAM ProNM, Carl Zeiss Meditec, Inc., Dublin, CA), automated static perimetry (Octopus 101, G1 program, Interzeag, Bern-Köniz, Switzerland) and/or manual kinetic perimetry (Goldmann Perimeter, Haag-Streit, Bern, Switzerland) and assessment of the retinal nerve fibre layer (OCT Stratus 3000, Carl Zeiss Meditec, Inc.). Inclusion criteria for control subjects were best-corrected visual acuity of 20/20 or better, normal colour vision and normal visual field without evidence of retinal or optic nerve pathology. None of the control subjects had a previous or current history of ocular disease.
None of the patients or control subjects had a systemic condition or medication use that was known to affect efferent pupillary function. Both eyes of patients and control subjects were tested independently but only the pupil response of one eye was included for further analysis. The eye with the worse visual function was selected in patients and the right eye was selected in control subjects.
We used a previously described apparatus (ColorDome Ganzfeld ERG Diagnosys, Lowell, MA, USA) and protocol that presented an equiluminant red (640 ± 10 nm) and blue (467 ± 17 nm) light stimulus in a stair step fashion at three different intensities (1, 10 and 100 cd/m2) for a duration of 13 seconds (s) per intensity step.9 The untested eye was occluded with a patch and the tested eye was monitored continuously by a dual-channel binocular eye frame pupillometer worn by the subject (Arrington Research, Scottsdale, AZ, USA).
The subjects remained several minutes under low mesopic light before testing. Pupil recording began 13 s before the first 1 cd/m2 light stimulus and continued after light stimulus termination for 60 s in darkness (0 cd/m2). Subjects were asked to look straight ahead, as if gating at distance, in order to minimize accommodation.
Main outcome measures:
The baseline pupil size (BS) was defined as the mean size during the 13 s before the first light stimulus and assigned a value of 1.0. Pupil size thereafter was reported as a ratio relative to the BS and referred to as relative pupil size (RPS).
From the pupil response curve, we arbitrarily defined the first 10 s following light termination as the ‘early’ re-dilation phase and the period between 10 and 30 s after light termination as the ‘late’ re-dilation phase. The following parameters were quantified:
1Pupillary re-dilation time. This was the duration of time (in seconds) after light stimulus termination needed to recover 90% of BS.
2Pupil size during early and late phases of re-dilation. This was the mean RPS between 0 to 10 s and 10 to 30 s, respectively, after light termination.
3Differential pupillary re-dilation time. This is re-dilation time following blue 100 cd/m2 light minus re-dilation time following red 100 cd/m2 light.
We also assessed the light stimulus phase of the pupil response curve in order to confirm that pupil contraction responses obtained in our patients to stimuli weighted to favour rods and cones were similar to those reported in a previous study using the same testing protocol. The contraction amplitude was calculated and reported as the mean percent pupil contraction from BS between 1 s and 3 s after the light stimulus onset and calculated for each light intensity step.
All pupillographic data were analysed off-line using EXCEL spreadsheet (Microsoft Excel 2003 Software). A customized filter was applied to remove most of the blink artefacts. The remaining blink artefacts, operationally defined as an abrupt, spontaneous, brief (<250 ms) and large change in pupil size (increase or decrease >50%) followed by a rapid return to the previous pupil size, were manually removed. The pupil response curve was plotted as the RPS as a function of time using Kaleidagraph 4.0 graphing software (Synergy Software, Reading, PA, USA) in order to visualize the dynamics of the pupil movement in response to the stimulus paradigm. A 5% weighted smoothing function was applied to the graph of the tracings to reduce non-physiological variations in the pupil tracing such as eye movements, which could cause artifactual changes in pupil shape and thus add additional noise in the recording.
The transient pupil response to red 100 cd/m2 and to blue 1 cd/m2 light stimulation and the sustained pupil response to blue 100 cd/m2 light stimulation were selected for further analysis as these stimulus conditions have been previously described to be weighted to cone, rod and melanopsin contribution, respectively, to the afferent pupillomotor input.9 The pupil responses for each of these stimulus conditions for patients with outer retinal disease and patients with optic neuropathy were compared with the pupil responses of the control subjects using t-test for normally distributed data or Mann–Whitney test for non parametric data. The median pupil contraction to each of these stimulus conditions was compared between all subject groups using anova one-way for normally distributed data or Kruskall-Wallis test for non-parametric data (GraphPadPrism Logiciel GraphPad Inc., La Jolla, CA, USA). Test for normality was performed using the Shapiro–Wilk test. Statistical significance was set at P < 0.05.
Normative data for full-field ERG (Roland Consult, Brandenburg, Germany) had been previously established in our laboratory. The patients with outer retinal disease were arbitrary divided in two subgroups based on their ERG combined response (scotopic b-wave amplitude with 0 dB stimulus). Group 1 were patients whose ERG combined response fell below 30% of the lower limit of normal, and Group 2 were those who had a combined response between 30 and 80% of the lower limit value.
A correlation analysis between the ERGs combined response and measures of pupillary re-dilation (re-dilation time and size after light termination) was performed using Pearson coefficient. An anova was performed comparing these pupil parameters between the control group and both subgroups of patients with outer retinal disease.
The Goldmann visual field (GVF) was quantified by a manual grid scoring system.13,14 A transparent grid template made up of 100 dots was overlaid on the visual field of the patient. Dots that fell within (but not on or outside) the I4e isopter were counted. The maximal score was 100. The GVF scores of the patients with outer retinal disease were ranked and correlated to the measures of pupillary re-dilation (re-dilation time and size after light termination) using Pearson correlation analysis.
Twenty-three patients with outer retinal disease, 13 patients with optic neuropathy and 14 control subjects with normal eyes were included for the study. The 23 patients with outer retinal disease were 13 women and 10 men with median age 40 years (range from 8 to 79 years). Diagnoses included retinitis pigmentosa in 14 patients, Leber congenital amaurosis in six patients, corneoretinal dystrophy of Bietti in one patient, cone-rod dystrophy in one patient and Stargardt disease in one patient. One patient with retinitis pigmentosa had mild sensorineural hearing loss and was found to harbour the USH2A mutation. No patient had any associated systemic condition. Twenty-one patients demonstrated predominantly rod dysfunction on ERG and two patients (cone-rod dystrophy and Stargardt disease) had predominantly cone dysfunction. Among the 13 patients with an optic nerve disorder (four women and nine men, median age 64 with range from 27 to 72), 11 had ischemic optic neuropathy and two had a compressive lesion of the optic nerve. In the control group, there were nine women and five men (median age 30 with range from 19 to 64).
The pupil contraction amplitudes of patients with outer retinal disease were not significantly different from that of patients with optic neuropathy (median = 25% and 31% for blue 1 cd/m2, respectively, and median = 40 and 39% for red 100 cd/m2, respectively, P > 0.05). The contraction amplitudes of both patient groups were, however, significantly smaller compared with the controls (see Table 1).
Table 1. Summary of the pupil contraction (in percent relative to baseline) to the dim blue light stimulation (rod-weighted response) and the bright red light stimulation (cone-weighted response) in the three groups of subjects (control subjects, patients with outer retinal disease (ORD) and patients with optic neuropathy (ON)
Patients with ORD
Patients with ON
Blue light stimulation at 1 cd/m2
Median = 50% Range: 36–58
Median = 25% Range: 6–45
Median = 31% Range: 5–42
P < 0.0001
P = 0.001
P = 0.67
Red light stimulation at 100 cd/m2
Median = 42% Range: 29–47
Median = 40% Range: 17–54
Median = 39% Range: 6–48
P = 0.008
P = 0.002
P = 0.26
The pupillary re-dilation was assessed following termination of the brightest stimulus intensity (100 cd/m2). Qualitative assessment of the pupil response curves showed that pupillary re-dilation after the bright red and bright blue light stimulus was prompt and rapid in control subjects and patients with optic neuropathy (Fig. 1). Specifically, there was no observable difference in pupillary dynamics during the dilation phase of the pupillogram between the blue or red light stimulus conditions. However, among the patients with outer retinal disease, there was an observable delay in pupillary re-dilation after termination of the bright blue light stimulus but not after the red light stimulus (Fig. 1). An illustrative case example of this differential pupillary re-dilation to red and blue light is shown in Figure 2 (top).
The pupillary re-dilation curves were nearly identical among control subjects and patients with optic neuropathy. The pupillary re-dilation times for the red light and blue light conditions were not significantly different between these two groups (P > 0.05). The median differential re-dilation time was 5.5 s with a range of 0 to 17 s for controls subjects and 5.0 s, range 0–11 s for patients with optic neuropathy. In comparison, the differential re-dilation time was significantly greater in patients with outer retinal disease (median = 28.0 s with range from 1 to 57 s, P < 0.0001 compared with controls). Two patients with outer retinal disease did not demonstrate a differential pupil re-dilation time that was greater than the median value of the control group. These two patients (both with a diagnosis of retinitis pigmentosa) had a differential re-dilation time of 1 and 3 s. One patient had an abnormal but still recordable ERG with acuity of 0.25.The other patient had a non-recordable scotopic and photopic ERG and very poor Goldmann field score of 3, though acuity was preserved at 0.8. This patient's pupillogram is shown in Figure 2 (bottom image).
The quantified parameters of pupil re-dilation (re-dilation time and size after light termination) for patients with outer retinal disease, patients with optic neuropathy and control subjects are summarized in Table 2. The time to recover 90% of BS was significantly longer after bright blue light stimulation in patients with outer retinal disease compared with the control group (34 s for patients vs. 16 s for control subjects; P < 0.05). The re-dilation time after bright red light stimulation was not significantly different between these two groups (5 s for patients vs. 11 s for controls; P > 0.05). Patients with optic neuropathy showed a slightly faster recovery time compared with control subjects for both blue and red light stimulus conditions but the difference was not significant.
Table 2. Summary of quantified parameters of pupillary re-dilation following termination of a red and blue light stimulus in three groups of subjects (control subjects, patients with outer retinal disease (ORD) and patients with optic (ON)
Measures of Pupillary Re-Dilation
Patients with ORD
Patients with ON
Value that falls outside the 5th or 95th percentile limit. BS, baseline pupil size; RPS, relative pupil size.
RED light stimulus condition
Time to recover 90% of BS
Median = 11 s 95th percentile = 17 s
Median = 5 s
Median = 4 s
Mean RPS between 0 and 10 s after light termination
Median = 0.82 5th percentile = 75
Median = 0.89
Median = 0.90
Mean RPS between 10 and 30 s after light termination
The mean RPS during the early and late phase after bright blue light stimulation was significantly smaller in patients with outer retinal disease compared with the control group (median = 0.58 s and 0.78 s for patients vs. 0.70 s and 0.90 s for controls, respectively; P < 0.05). In contrast, after red light stimulation, the mean RPS for both periods of dilation was not significantly different between patients with outer retinal disease compared with the control group (median = 0.89 s between 0 and 10 s and 0.94 s from 10 s to 30 s vs. 0.82 s and 0.94 s, respectively, for controls; P > 0.05). Patients with optic neuropathy had a mean RPS during the early and late phase of re-dilation that was significantly larger than that of control subjects, for both blue and red light stimulus conditions, and this finding was consistent with the faster re-dilation time seen in this group of patients.
Ten patients with outer retinal disease had a non-recordable ERG. Eight of these patients had a quantifiable pupil response to red and blue light stimulation at all intensities; 2 patients did not show a pupil contraction to the blue 1 cd/m2 light stimulation. Nine patients demonstrated abnormally delayed pupillary re-dilation after termination of the bright blue light stimulus. There was no significant correlation between the ERG and any of the quantified parameters of pupillary re-dilation.
The 23 patients with outer retinal disease were divided into two groups: Group 1 (n = 17) and Group 2 (n = 6) based on the ERG combined response (see Data Analysis, under Methods). Following termination of the bright blue light stimulus, Group 1 exhibited a longer time to recover to 90% of BS (37 s) and had a lower mean RPS (0.58 from 0 and 10 s and 0.75 from 10 to 30 s after stimulus offset) compared with control subjects (time = 16 s, mean RPS = 0.7 and 0.9, respectively). These differences between Group 1 and the control group were statistically significant (P < 0.05). There was no significant difference in any of these parameters between Group 2 and the control group.
The GVF score in the patients with outer retinal disease ranged from 0 to 27. Sixteen patients had a GVF score of 0. There was no significant correlation between the GVF score and time to recover 90% of BS, mean RPS from 0 to10 s after stimulus light termination and mean RPS from 10 to 30 s after stimulus light termination (r = 0.28, −0.27 and −0.23, respectively, and P > 0.05).
To blue and red light stimuli presented under conditions pre-selected to favour rod and cone contribution to the pupil light reflex, patients with outer retinal disease had significantly reduced pupil contractions compared with control subjects with normal eyes. These findings are not unexpected and are consistent with those from a previous study using coloured light stimuli to assess outer photoreceptor function in patients with retinitis pigmentosa.10 We found that the light responses were similarly reduced in the patients with optic neuropathy, suggesting that pupil contraction amplitude to dim blue light (1 cd/m2) and to bright red light (100 cd/m2) is not a response parameter that distinguishes between outer retinal disease and optic neuropathy.
In contrast, the pupillary re-dilation dynamics after light termination was not similar between the two patient groups. The patients with optic neuropathy showed a prompt and rapid re-dilation after offset of bright blue as well as bright red light, just like that seen among the control subjects. The red and blue light re-dilation response curves were nearly superimposed on each other in the control group and in the patients with optic neuropathy, resulting in a negligible differential re-dilation time (difference in time to recover baseline size following red vs. blue light). In the patients with outer retinal disease, however, after bright blue light offset the pupil stayed smaller in size for a variable period time and overall, pupillary re-dilation was notably slower with significantly longer time to recover baseline size, compared with that after bright red light offset. The differential re-dilation time in patients with outer retinal disease was significantly greater compared with controls (28.0 s vs. 5.5 s, respectively). This was a very distinctive finding as none of the control subjects or patients with optic neuropathy had a differential re-dilation time that reached the median time of the patients with outer retinal disease.
It is important to recognize that delayed pupillary re-dilation after exposure to blue light has no specific significance by itself. The finding must be related to the light stimulus and to a control group. For example, it has been shown that a bright enough light can evoke pupillary contraction that persists after light termination and thus cause delayed pupillary re-dilation in normal eyes.15,16 In the study by Kankipati et al., stimulation of normal eyes by a short wavelength (blue) light of high energy (470 nm, 13 log quanta/cm2/s for 10 s) evoked a strong and sustained pupil contraction.17 This prolonged state of pupillary contraction after blue light termination was attributed to preferential activation of melanopsin and a predominant effect of intrinsic signalling on the MGC activity, as it was not present after exposure to an equivalent long wavelength (623 nm, red) light stimulus. As such, Kankipati et al. used the post-illumination pupil contraction as a means to measure MGC activity and to quantify changes in disease states that primarily affect the ganglion cells.17
With our testing protocol, the bright blue bright light stimulus does not evoke a pupil contraction that persists after light termination. This has been reported in previous studies that introduced this testing protocol and likewise, it was not observed in any of our control subjects.9,10 The main reason for this is simply insufficient light energy in our bright blue light stimulus to strongly and preferentially activate melanopsin (Dr Gamlin, pers. comm., 2006 Tokyo, Japan). If the pupil response obtained with our bright blue light stimulus reflects a signal of mixed photoreceptive input, for example, melanopsin plus S/M cones and rods, then the re-dilation dynamics is also influenced by the outer photoreceptors that promptly transmit an OFF signal to the ganglion cells at the moment of light offset.18 Inhibition of the Edinger–Wesphal nucleus combined with activation of the descending sympathetic pathways leads to the characteristic prompt and rapid pupillary dilation that is seen with the range of light stimuli used in clinical setting (lights having low to moderately bright intensities).
Our finding of delayed pupillary re-dilation in patients with outer retinal disease supports the notion that our bright blue light stimuli, when presented to normal eyes, results in activating outer and inner retinal photoreceptors. In the patients with outer retinal disease, we propose that the loss of rods and cones causes of the proportion of photoreceptive input to be pathologically altered in favour of melanopsin. To light stimulus conditions (moderately bright blue light) that do not normally evoke a post-illumination pupil contraction but are weighted to activate melanopsin, the greater contribution of melanopsin and the greater influence of intrinsic signalling on the overall MGC activity will be unmasked in eyes with outer retinal disease. In addition, the loss of OFF-signalling of rods and cones on MGCs can also contribute to a delay and slowing of pupillary re-dilation. The pupil re-dilation to bright red light would not be expected to change in patients with outer retinal disease as the long wavelength heavily shifts photoreceptor activation in favour of M/L cones and is not weighted enough in favour of melanopsin.
In order to compare the light OFF response between patients and controls, we defined and quantified several measures of the pupillary re-dilation portion of the pupillogram. Twenty-one of 23 patients with outer retinal disease had abnormal pupillary re-dilation parameters and a differential re-dilation time that was outside the limits of the control group. Although patients with the most severe outer retinal disease, for example, unrecordable ERG and worst visual function, tended to have the most abnormal pupil re-dilation parameters in the blue light condition, no correlation could be found between any of the re-dilation parameters and GVF or ERG.
One possible reason for the lack of correlation is simply that many of the patients with outer retinal disease in our study were already in a very advanced stage of their disease. Ten of 23 patients had an unrecordable ERG that had been flat for a long time prior to study entry. In addition, the range of GVF scores was very limited (0 to 27 on a scale of 0 to 100, 16 patients with outer retinal disease had a GVF score of zero). Finally, in some patients, the pupillometry was not performed at the same time as the ERG examination, and this mismatch in timing of the tests may have added to the lack of correlation between pupillographic and clinical results. In a post-analysis inter-eye comparison among patients with outer retinal disease, the majority (15 of 22 patients, one patient was monocular) demonstrated greater abnormality of pupillary re-dilation parameters in the eye with worse visual field and ERG. Further studies will be needed to clarify whether or not a correlation truly exists.
Despite the severe loss of electrophysiological and subjective visual function, all patients had recordable and quantifiable pupil responses to red and blue light. Only two patients had an unrecordable pupil response to only the dim blue light stimulus (1 cd/m2), and their pupil responses to 10 and 100 cd/m2 intensity were easily recordable. This retained capacity for the severely diseased retina to evoke a pupil light reflex suggests that coloured light pupillometry may be a more sensitive method to assess low levels of rod and cone functioning compared with standard ERG or GVF.
In conclusion, we have recorded and quantified pupillary re-dilation after termination of a high intensity, coloured light stimulus in control subjects with normal eyes and patients with neuroretinal visual loss. Under conditions of light presentation used in this study, only patients with outer retinal disease demonstrated a quantifiable difference in pupillary re-dilation to blue and red light conditions due to a slowing of re-dilation after bright blue light exposure that significantly delays the recovery time back to baseline size. The control subjects with normal eyes and patients with optic nerve disease did not show any significant difference in pupillary re-dilation under red and blue light conditions. In fact, the re-dilation dynamics of patients with optic neuropathy was not distinguishable from that of the control subjects. Ongoing studies are anticipated to better define the sensitivity of using pupillary re-dilation as a parameter to differentiate outer retinal disease from optic neuropathy as well as to monitor disease progression.
The authors thank Yann Leuba for his technical assistance for the graphic display and layout used in the figures.