top of page

Prosody and the Fluency Rubrics: What WCPM Does Not Capture

Writer: Andrea Chernin
Andrea Chernin
3 days ago
19 min read

Every published definition of reading fluency has three parts. Speed, accuracy, and expression. Every number that reaches your progress-monitoring graph has two. A reading prosody rubric is the only assessment instrument most practitioners own that scores the third part at all, and the four rubrics in common use disagree about the number of dimensions, the range of the scale, and where the line between fluent and not-fluent falls.


This is not a gap in the research. It is a gap between the definition the field agreed on and the instruments the field actually administers.

Every definition includes expression. Every score leaves it out.

The Report of the National Reading Panel defines it in one clause. Fluency is "the ability to read a text quickly, accurately, and with proper expression." Its own executive summary opens the same way: "Fluent readers can read text with speed, accuracy, and proper expression."


Eighteen years later the National Center for Education Statistics used almost identical wording. The 2018 NAEP Oral Reading Fluency Study states that "oral reading fluency is defined as the ability to read text aloud with speed, accuracy, and proper expression."


Three components in the definition. Two components on the report.

The field has noticed. Kuhn, Schwanenflugel and Meisinger made the argument directly in Reading Research Quarterly in 2010, writing that practice had settled on "a relatively narrow definition of reading fluency, one that emphasizes automatic word recognition."


Their paper examines four competing definitions against accuracy, automaticity and prosody, and considers what each implies for assessment. The full citation is Kuhn, M. R., Schwanenflugel, P. J., and Meisinger, E. B. (2010), Reading Research Quarterly, 45(2), 230 to 251, doi 10.1598/RRQ.45.2.4.


Sixteen years after that paper, nothing on the standard curriculum-based measurement form has changed.

What your WCPM measures actually count

This is worth checking rather than assuming, because the measures differ from each other and none of them scores expression. If you use the words-correct-per-minute norm tables, you are reading a number built from two of the three components.


DIBELS 8th Edition


The DIBELS 8 Administration and Scoring Guide, published by the University of Oregon, is explicit about the Oral Reading Fluency score. "The final score is the number of words read correctly (and self-corrected) within one minute."


Errors are defined mechanically. "Words omitted, substituted, and hesitations of more than three seconds are scored as errors." A self-correction inside three seconds counts as accurate.


Nothing in the ORF scoring rules addresses phrasing, pausing, stress or intonation. The only timing rule in the measure is the three-second hesitation threshold, which is an error rule rather than a prosody judgement. The guide also records that Retell Fluency was removed from DIBELS 8, so the one subtest that asked the student to do something other than read fast and accurately is no longer there.


Acadience Reading K to 6


The Acadience Reading ORF and Retell page lists exactly four reported scores. Median words correct per minute, median errors per minute, median correct words in the Retell, and median Quality of Response for the Retell.


Note what is not on that list. There is no accuracy percentage, and there is no expression score. Errors are counted the same mechanical way, including "words read incorrectly, substitutions, skipped words, hesitations of more than 3 seconds." The Retell Quality of Response rates how well the student retold the passage, not how the reading sounded. The Acadience Reading technical manual is the place to read the psychometrics behind those four scores.


GORT-5


This is the one that deserves a second look, because the Gray Oral Reading Tests, Fifth Edition reports a score named Fluency, and that score contains no prosody at all.


Pearson's product page lists the scores as Rate, Accuracy, Fluency, Comprehension and the Oral Reading Index. Fluency is derived by combining Rate and Accuracy. The Oral Reading Index is a composite of Fluency and Comprehension, reported as a standard score with a mean of 100 and a standard deviation of 15.


The authors are J. Lee Wiederholt and Brian R. Bryant, the age range runs from 6:0 to 23:11, and purchase sits at qualification Level B.


When a technical manual labels Rate plus Accuracy as "Fluency," the word on the report page is doing less work than the word in the definition.

That is not a defect in GORT-5. It is a naming convention the whole field uses, and it is the reason a parent or a teacher reading your report may believe expression was assessed when it was not.

The four rubrics that do score prosody

There are four instruments in wide use that put a number on expression. They are not interchangeable, and the differences are not cosmetic.


The 1992 NAEP oral reading fluency scale


This is the original, and it is still the one most often reproduced in state guidance. NAEP's Listening to Children Read Aloud fact sheet, published in August 1995, reports the scale and the fourth-grade results from the 1992 Integrated Reading Performance Record. The sample was 1,136 fourth-grade students.


The scale has four levels, assigned holistically to a whole passage reading.


  • Level 4: "Reads primarily in larger, meaningful phrase groups." Regressions and repetitions may be present but do not detract from the structure of the story. "Preservation of the author's syntax is consistent." "Some or most of the story is read with expressive interpretation."

  • Level 3: "Reads primarily in three- or four-word phrase groups." The majority of phrasing "seems appropriate and preserves the syntax of the author." Critically, "Little or no expressive interpretation is present."

  • Level 2: "Reads primarily in two-word phrases with some three- or four-word groupings." Groupings "may seem awkward and unrelated to larger context of sentence or passage."

  • Level 1: "Reads primarily word-by-word."


The distribution NAEP found in 1992 was 13 percent at Level 4, 42 percent at Level 3, 37 percent at Level 2, and 7 percent at Level 1. The full synthesis of the 1992 data by Pinnell and colleagues is the primary document behind every adapted version you will encounter.


Two things about this scale are routinely lost in reproduction. First, Level 3 explicitly requires little or no expressive interpretation, so a reader can reach the second-highest level without any expression at all. Second, the cut is stated: Colorado's guidance puts it plainly, that "Students at levels 3 and 4 are generally considered to be fluent, and those at levels 1 and 2 non-fluent."


The 2018 NAEP expression scale


NAEP ran the study again in 2018, with over 1,800 fourth-graders from 180 public schools, and it did not use the 1992 scale.


The 2018 study report scores expression on a 0 to 5 scale, with three additional non-scale codes. The levels are named, and the names describe a progression through linguistic units rather than through phrase length.


  • 0, Insufficient sample: fewer than 12 words read aloud correctly.

  • 1, Word by word: less than a quarter of words read with appropriate expression, mostly monotone.

  • 2, Local grouping: between a quarter and half of words read with appropriate expression, focused on local word groups.

  • 3, Phrase and clause: more than half of words read with appropriate expression, expressing words, phrases, clauses and a few sentences, but not the passage as a whole.

  • 4, Sentence prosody: more than three-quarters of words read with appropriate expression, occasionally inconsistent but not monotone, with a rate of at least 55 words per minute.

  • 5, Passage expression: at least 90 percent of words read with appropriate expression, consistently expressing sentence, paragraph and passage structure, with a rate of at least 80 words per minute.


Codes 8 and 9 mark a silent reader and an anomaly and are treated as missing. The highlights report and the study landing page carry the summary figures.


The Multidimensional Fluency Scale


This is the rubric most educational therapists have in a folder somewhere, and it is the only one of the four that produces a numeric total.


Rasinski's Assessing Reading Fluency, published in 2004 by Pacific Resources for Education and Learning in Honolulu, reproduces it. The scale has four dimensions: Expression and Volume, Phrasing, Smoothness, and Pace. Each is scored 1 to 4. The document states that "Scores range from 4 to 16."


It also states a decision rule, which none of the other three rubrics does: "Generally, scores below 8 indicate that fluency may be a concern. Scores of 8 or above indicate that the student is making good progress in fluency."


Hold onto that number. It matters in a moment.


The same document carries the accuracy and rate reference points practitioners usually take from it. Accuracy bands are given as 97 to 100 percent for the independent level, 90 to 96 percent for the instructional level, and below 90 percent for the frustration level.


Grade-level target rate ranges are given for fall, winter and spring, with the note that "Readers who perform at or near these target norms should be considered as progressing adequately in automaticity."


The Fountas and Pinnell fluency scale


The fourth is the one bundled with a widely used benchmark assessment system, and it is the one whose shape is hardest to pin down, because it appears in more than one form.


The version on Heinemann's own site, inside the Benchmark Assessment System recording form, records a single Fluency Score of 0, 1, 2 or 3. The scoring key folds phrasing, expressive interpretation, pausing, stress and rate into one band per level.


Level 0 is "Reads primarily word-by-word with occasional but infrequent or inappropriate phrasing; no smooth or expressive interpretation, irregular pausing, and no attention to author's meaning or punctuation; no stress or inappropriate stress, and slow rate."


Level 3 is "Reads primarily in larger, meaningful phrases or word groups; mostly smooth, expressive interpretation and pausing guided by author's meaning and punctuation; appropriate stress and rate with only a few slowdowns." The form carries a 2011 and 2008 copyright to Irene C. Fountas and Gay Su Pinnell.


If that Level 3 language sounds familiar, it should. It is close to the 1992 NAEP Level 4 wording, compressed.


A six-dimension version also circulates, scoring Pausing, Phrasing, Stress, Intonation and Rate separately plus Integration as an overall judgement, where "Integration involves the way a reader consistently and evenly orchestrates rate, phrasing, pausing, intonation, and stress."


The copies of the system toolkit that carry it are district-hosted rather than publisher-hosted, and the publisher's own toolkit pages are not readable by an automated fetcher, so treat the six-dimension form as real but verify the edition you own against your own kit rather than against this article.

One construct, four scales, no conversion

Put the four side by side and the problem is visible without any statistics.


  • NAEP 1992: one holistic judgement over the whole passage, four levels numbered 1 to 4, with Levels 1 and 2 stated to be non-fluent.

  • NAEP 2018: one holistic judgement over the whole passage, six scale points numbered 0 to 5, with no cut score stated.

  • Multidimensional Fluency Scale: four dimensions scored separately, each 1 to 4, for a total range of 4 to 16, with totals below 8 stated to be a concern.

  • Fountas and Pinnell recording form: one holistic judgement over the whole passage, four levels numbered 0 to 3, with no cut score stated.


There is no published crosswalk between any two of these. A Level 3 on the 1992 scale, a 3 on the 2018 scale, a 12 on the Multidimensional Fluency Scale and a 2 on the Fountas and Pinnell form are four different statements about four different things, and none of them converts into another.


This is the same structural problem that turns up with reading level systems, where four scales coexist with no defensible conversion, and with the four score metrics on a psychoeducational report. The difference is that nobody publishes a prosody conversion chart, which is at least honest.

The two NAEP scales measure different things, and the second one knows it

It would be easy to read the 2018 study as a continuation of the 1992 one. It is not, and the report does not claim it is.


The 1992 scale is a phrasing scale with expression attached at the top. Read the level descriptors again and you will see that Levels 1, 2 and 3 are distinguished almost entirely by phrase length, and expression enters only at Level 4. A reader who phrases in four-word groups and reads completely flat lands at Level 3.


The 2018 scale is a proportion-of-words scale. Each level is defined by what fraction of words were read with appropriate expression, with the linguistic unit described alongside. Expression is present at every level above 1, in increasing amounts.


Those are not two versions of one scale. They are two different theories of what prosody is, thirty years apart, from the same assessment programme. If you have reproduced the 1992 rubric in your intake paperwork because it is the one that circulates, you are using NAEP's older theory of the construct.

The 2018 rubric says rate is not part of expression, and then uses rate

This is the detail most worth knowing, and it is sitting in plain sight in the report.


The 2018 study states that speaking rate within a normal range, word accuracy and pronunciation are not part of the expression score. Substitutions and skipped words do not count against expression if the meaning is reasonably conveyed. That is a deliberate design choice, and the point of it is to produce a measure of expression that is not contaminated by speed.


Then Level 4 requires a rate of at least 55 words per minute, and Level 5 requires at least 80.


Both statements are in the same rubric. The practical consequence for a practitioner is specific and it is the one thing to carry away from this section: a student who reads with genuinely good expression at 45 words per minute cannot score above Level 3 on the 2018 scale. The ceiling is set by rate, on a measure that excludes rate by definition.


That is not a reason to discard the scale. It is a reason to stop treating an expression score as independent evidence when you are already graphing words correct per minute, because above Level 3 the two are mechanically linked.

What NAEP's own agreement figures say

The 2018 study reports how well two independent human scorers agreed when scoring expression on that 0 to 5 scale. The figures are published and they are modest.


Exact agreement between two independent scorings was 58 percent, with a further 39 percent agreeing within one level.

The report also states the quality thresholds NAEP applies to its own scoring: exact agreement below 61 percent triggers a mild concern flag, and below 57 percent a greater concern. Fifty-eight percent sits inside the mild-concern band.


Read that carefully, because it is easy to over-read in either direction. It does not mean the scale is worthless. Ninety-seven percent of the time two trained scorers landed within one level of each other, which is reasonable for a holistic judgement.


It does mean that a single expression level, from a single scorer, on a single passage, is a soft number, and it should not be the thing a programme decision turns on. The same logic applies to the standard error around any single score you report.


And these are trained NAEP scorers, working from recordings, on a rubric with written anchors, under a quality-control regime. Your own scoring of a student you know well, live, in session, is unlikely to be more reliable than that.

The reliability problem has a measured size, and one dimension fails

Someone has measured this properly on the Multidimensional Fluency Scale, and the result is specific enough to change practice.


Duong, Mostow and Sitaram, writing in ACM Transactions on Speech and Language Processing in 2011, had two judges score oral reading on the four-dimension rubric and reported intraclass correlations at two grain sizes. Their paper, Two Methods for Assessing Oral Reading Prosody, is the clearest published account of what the rubric can and cannot carry.


At the student level, averaging over 20 sentences each for 10 students, the intraclass correlations were 0.798 for expression, 0.681 for phrasing, 0.721 for pace, and 0.685 overall. Smoothness came in at 0.283.


At the sentence level, across 200 sentences, everything dropped: 0.593 for expression, 0.535 for phrasing, 0.398 for smoothness, 0.578 for pace, 0.638 overall. Within-one-point agreement ranged from 84 percent for smoothness to 94 percent for expression.


The authors drew two conclusions and both are practical. A single sentence is too small a sample to rate reliably on this rubric. And their human ratings were too unreliable to serve as a gold standard for validating an automated measure, which is a strong thing for authors to say about their own data.


Three things follow for a caseload.


  • Score prosody over a substantial sample of connected text, not a line or two. The published reliability at sentence level does not support a per-sentence judgement.

  • Treat Smoothness as the weakest of the four dimensions. An intraclass correlation of 0.283 between two judges at the student level means that dimension is carrying very little signal, and the authors attribute part of it to the difficulty of observing pauses on playback.

  • Do not report a change of one point on a single dimension as progress. It is inside the noise.

Where the rubrics put the line, and why they disagree

Now return to that Multidimensional Fluency Scale cut score, because placing it next to the NAEP cut produces a contradiction nobody has resolved.


The Multidimensional Fluency Scale says scores of 8 or above indicate good progress. Eight out of 16 across four dimensions is an average of exactly 2 per dimension.


The NAEP 1992 scale says Levels 1 and 2 are non-fluent.


So a reader scoring a flat 2 on all four Multidimensional dimensions is "making good progress in fluency" on one rubric and "non-fluent" on the other. The descriptors at those levels are not far apart either: both describe two-word to four-word phrasing with little expression.


That is a genuine disagreement about where adequate ends, between two of the four most widely reproduced prosody rubrics in American practice. It is not a rounding difference, and neither document acknowledges the other.


If you are going to use a cut score, use one rubric's cut with that rubric's descriptors and say which in your notes. Mixing them produces a number nobody can interpret, which is the same failure mode as reading a composite whose manual told you not to.

The Fountas and Pinnell scale exists in at least three shapes

This is worth flagging separately because it shows up in the research literature in a form that does not match the publisher's own recording form.


The recording form on Heinemann's site records one Fluency Score, 0 to 3.


District-hosted copies of the system toolkit describe six dimensions scored 0 to 3, five of them specific and the sixth an integration judgement.


And a peer-reviewed study using the rubric, Sarris and Dimakos in Themes in Science and Technology Education, 8(1), 47 to 62, describes a "Six Dimensions Fluency Rubric" attributed to Fountas and Pinnell (2009) in which each dimension is scored 1 to 4, not 0 to 3.


Three experienced teachers scored each recording and their scores were averaged. That study measured prosody against acoustic features with Praat and reported correlations of negative 0.83 with voice breaks, 0.72 with phonation time ratio and 0.78 with pitch ratio on its control text, across 27 Greek-speaking students in grades 2 and 5.


One name, three scales. A 1 to 4 six-dimension version has a 6 to 24 range. A 0 to 3 six-dimension version has a 0 to 18 range. The single holistic version has a 0 to 3 range.


If a report in a student's file says "Fountas and Pinnell fluency: 3," you cannot tell from that sentence whether it means the top of a four-point holistic scale or the bottom third of a 24-point one. Ask which form was used, and write which form you used.

What a prosody score can and cannot support you claiming

This is the part where the boundary matters, so here it is stated plainly rather than implied.


The 2018 NAEP study describes a "strong and consistent relationship" between expressivity and NAEP reading performance. Average expression declined steadily from Advanced to Proficient to Basic and continued to decline through the below-Basic subgroups, with the overall average near Level 4 and the lowest below-Basic subgroup falling below Level 3.


That is a cross-sectional association in a nationally representative sample, measured at one point in time. It is good evidence that readers who comprehend well also read aloud expressively.


It is not evidence that teaching prosody raises comprehension, and you may not draw that conclusion from it. The study did not intervene, did not randomise, and did not follow anyone over time. Nothing in it tells you what would happen to a student's comprehension if you spent eight weeks on phrasing and expression.


The automated-measurement work points the same way and has the same limit. Duong and colleagues report cross-validated within-grade correlations of 0.763 between a template-based prosody model and posttest comprehension, and 0.504 with comprehension gain, across 55 students. Those are prediction figures from a modelling study, not intervention effects. A measure that predicts an outcome is not thereby a treatment for it.

What the published instructional guidance actually recommends

If you want an evidence-rated recommendation that touches this ground, the nearest one is not about prosody at all.


The IES practice guide Foundational Skills to Support Reading for Understanding in Kindergarten Through 3rd Grade carries four recommendations. The fourth is to "Ensure that each student reads connected text every day to support reading accuracy, fluency, and comprehension." Its summary document records that this recommendation was assigned a moderate level of evidence.


The reason for moderate rather than strong is stated. The What Works Clearinghouse identified 22 studies; 18 showed positive effects on word reading, oral reading accuracy, oral reading fluency or comprehension, but eight of those showed an effect in only one domain, three showed no effects at all, and one showed a negative effect on one outcome. The panel's own summary calls the evidence "inconsistent."


The guide's language about how connected text should be read is about accuracy, appropriate phrasing and comprehension, with oral reading practice and constructive feedback. That is closer to a repeated reading protocol than to a prosody curriculum, and it is worth noticing that the recommendation names phrasing without recommending that you score it.

Prosody is not a domain on the tools charts

One practical consequence of all of the above is that you will not find prosody among the screening or progress-monitoring domains on the National Center on Intensive Intervention tools charts. The reading measures reviewed there are built on the metrics that can be counted reliably and frequently, which is what a progress-monitoring measure needs.


That is a defensible position for a tools chart to take. A measure with 58 percent exact inter-rater agreement is not a good weekly progress-monitoring probe, and what a full bubble on that chart actually requires is a specific psychometric standard that a holistic rubric will not meet.


It does mean that if your practice relies only on charted tools, expression is absent from your data by construction rather than by choice. Knowing which it is matters when a parent asks why the graph went up and the reading still sounds flat.

Using a prosody rubric in a practice without overreading it

Here is what the sources above support, and nothing beyond it.


  • Pick one rubric and keep it. Switching between the 1992 NAEP scale and the Multidimensional Fluency Scale mid-case makes your own records uninterpretable, and there is no conversion to fall back on.

  • Score from a substantial connected-text sample. The published reliability at sentence level does not support rating a line or two.

  • Record the rubric by name, edition and range every time. "Fluency 3" is not a score. "Fountas and Pinnell recording-form holistic fluency, 3 of 3" is.

  • Use it descriptively alongside the counted metrics, not as a second independent outcome. On the 2018 scale above Level 3 it is partly a rate measure anyway.

  • Do not set a goal on a single dimension. The dimension-level reliability does not carry it, and Smoothness in particular does not.

  • Keep the cut score with its own rubric. Below 8 of 16 is a Multidimensional Fluency Scale statement; Levels 1 and 2 are a NAEP statement. They disagree and you cannot average them.


Where expression is a genuine presenting concern, it belongs in your session notes as a described observation with the rubric named, and in your progress-monitoring decision rules as context rather than as the trigger.

What to write in the report

A defensible paragraph looks like this, and it is short.


Oral reading was sampled from connected text at an instructional level and rated on the Multidimensional Fluency Scale (Rasinski, 2004), which scores Expression and Volume, Phrasing, Smoothness and Pace from 1 to 4 for a total of 4 to 16. The total was 9, and the rubric treats 8 and above as adequate progress.

Rubric-based prosody ratings carry modest inter-rater reliability, and dimension-level scores should not be read as independent measures. The rating is reported as a description of how the reading sounded, alongside words correct per minute, and not as a separate outcome.

That wording names the instrument, the range, the cut, the source of the cut, and the limitation. It is short, and it will not come back to haunt you. It also avoids the trap of implying that the rubric and the words-correct figure are two independent pieces of evidence pointing the same way.


If you are reporting across norm-referenced and criterion-referenced instruments in the same document, say which the rubric is. A prosody rubric is criterion-referenced with no norm sample behind it, and readers of your report will often assume otherwise.

Where the published record runs out

Several things a practitioner would reasonably want to know are not publicly readable, and it is better to say so than to fill the gap.


  • The publisher's own current toolkit documentation for the Fountas and Pinnell fluency scale is not readable by an automated fetcher, so the six-dimension form is described here from district-hosted copies and from a peer-reviewed study that used it. The authoritative version is the one in your own kit. Open it and check the range before you report a score from it.

  • No inter-rater reliability figure is published for the 1992 NAEP four-level scale in the documents cited here. The Colorado guidance reproducing it reports none.

  • No published crosswalk exists between any two of the four rubrics compared above, and this article does not construct one.

  • The 2018 NAEP report gives expression results by achievement level in figures and appendix tables rather than as coefficients in its text, so no correlation between expression and NAEP reading score is quoted here.

  • Whether a prosody rubric score adds anything to a WCPM figure for an individual case, in terms of instructional decisions that turn out better, has not been established by any study cited here. The honest position is that it adds description, not decision power.


Nothing here is legal advice, and nothing here is a diagnostic recommendation. Which instruments you are qualified to administer and report is a matter for your own credential, your supervisor and the publisher's qualification requirements, and your obligations around test materials sit with test security.

Frequently asked questions

Does WCPM measure prosody at all?


No. Every measure checked here scores words read correctly in a fixed time, with errors defined mechanically. DIBELS 8 scores the number of words read correctly in one minute. Acadience reports median words correct, median errors, and two Retell scores. GORT-5 derives its Fluency score from Rate and Accuracy. None of the three scores phrasing, pausing, stress or intonation.


Which prosody rubric should I use?


The sources do not support naming a best one. They do support picking one and staying with it, because no crosswalk exists between any two of them. The Multidimensional Fluency Scale is the only one of the four that publishes a cut score, which makes it the easiest to report defensibly. The 2018 NAEP scale has published inter-rater agreement figures, which makes it the easiest to caveat honestly.


Why does the same student score differently on two prosody rubrics?


Because the rubrics measure different things on different scales. The 1992 NAEP scale is mostly a phrasing scale and admits expression only at its top level. The 2018 NAEP scale rates the proportion of words read with appropriate expression. The Multidimensional Fluency Scale sums four separate dimensions. The Fountas and Pinnell recording form compresses five features into one holistic band.


Can I use a prosody rubric for weekly progress monitoring?


The published reliability does not support it. NAEP's own two-scorer exact agreement on its 0 to 5 expression scale was 58 percent, and dimension-level intraclass correlations on the Multidimensional Fluency Scale fell as low as 0.283 for Smoothness at the student level. Prosody is not among the domains reviewed on the National Center on Intensive Intervention progress-monitoring chart.


Does improving prosody improve comprehension?


That has not been established by the sources here. The 2018 NAEP study found a strong and consistent association between expressivity and reading performance, but it was cross-sectional and did not intervene. The nearest evidence-rated recommendation, IES Recommendation 4 on reading connected text daily, carries a moderate rating and is about accuracy, fluency and comprehension rather than about prosody instruction.


What is a defensible way to report an expression rating?


Name the rubric, its edition, its dimensions and its range; give the score; give the cut score with its source; and state that rubric-based prosody ratings have modest inter-rater reliability and are offered as description rather than as an independent outcome measure.

If this is the kind of detail you want

Illuminate Educational Therapy Group publishes this material for practitioners: educational therapists, learning specialists and executive-function coaches who read technical manuals and want to know what an instrument actually scores before they report it. If you are not sure where your own role sits in that landscape, start with the practitioner scope map.


Join the educator list and new practitioner resources arrive as they are published.

Recent Posts

See All

Comments


bottom of page